Implementation and Evaluation of Personal Genetic Testing As Part of Genomics Analysis Courses in German Universities

doi:10.21203/rs.3.rs-2324352/v1

Purpose

Due to the increasing application of genome analysis and interpretation in medical disciplines, professionals require adequate education. Here, we present the implementation of personal genotyping as an educational tool in two genomics courses targeting Digital Health students at the Hasso Plattner Institute (HPI) and medical students at the Technical University of Munich (TUM).

Methods

We compared and evaluated the courses and the students’ perceptions on the course setup using questionnaires.

Results

During the course, students changed their attitudes towards genotyping (HPI: 79% [15 of 19], TUM: 47% [25 of 53]). Predominantly, students became more critical of personal genotyping (HPI: 73% [11 of 15], TUM: 72% [18 of 25]) and a majority of students stated that genetic analyses should not be allowed without genetic counseling (HPI: 79% [15 of 19], TUM: 70% [37 of 53]). Students found the personal genotyping component useful (HPI: 89% [17 of 19], TUM: 92% [49 of 53]) and recommended its inclusion in future courses (HPI: 95% [18 of 19], TUM: 98% [52 of 53]).

Conclusion

Students perceived the personal genotyping component as valuable in the described genomics courses. The implementation described here can serve as an example for future courses in Europe.

Decreasing genome sequencing costs, new technologies, and an improved knowledge of the human genome brought genomic research to a level of application in life sciences and medicine that exceeds expectations from three decades ago when the Human Genome Project was launched.¹ Translation of genomic research towards an application in clinical care is continuously progressing, with the aim of improving diagnosis, disease prevention, risk prediction, as well as drug efficacy and safety assessments.^2–6 An understanding of data science has become a requirement for genomic researchers.¹ In addition, the interest of European citizens in direct-to-consumer (DTC) genetic testing in a non-clinical setting is high, even though genetic education and understanding is low in the general population.⁷

To train a new generation of experts at the interface between computer science and medicine, the Digital Health (DH) Master’s program was initiated in 2018 at the Digital Engineering Faculty of the University of Potsdam, the Hasso Plattner Institute (HPI).⁸ In 2020, we introduced the Analyze Your Personal Genome (AYPG) course that incorporates voluntary personal genetic testing (PGT). Its aim is to enhance the understanding and to develop a sensitivity towards genetic testing by affecting the students on a personal level. The underlying rationale is based on self-determination theory: the students’ interest in the course content and their motivation for engaging in the course shifts from extrinsic motivation driven by grading to the intrinsic motivation of wanting to learn about their personal genomic background.⁹ In addition to increased motivation, we aimed for enhanced compassion with patients and individuals undergoing PGT. Because DH students have heterogeneous backgrounds, a particular challenge was to design a course that provides learning opportunities to everyone, while ensuring that students take informed decisions whether to undergo PGT, regardless of their previous knowledge.

In response to the rising application of clinical and DTC genetic testing, the necessity of a profound genetic education in medical training was already recognized more than a decade ago.¹⁰ At Tufts University and Stanford School of Medicine, discussions about introducing PGT into genomics courses to improve learning outcomes and motivation began in 2008 and 2009, respectively, and, after careful consideration, resulted in similar course concepts including voluntary genetic testing.^11,12 At the Icahn School of Medicine at Mount Sinai, New York, personal genome sequencing was first applied in medical training in 2012.¹³ Strategies and results of these three schools were summarized and discussed by Garber et al.¹⁴ Briefly, several schools aimed to improve the genomics education for healthcare providers by adding an innovative and engaging component to their curricula, such as personal genotyping or working with anonymized or donated genomic data. Enhanced educational outcomes of students participating in PGT were indicated in studies accompanying the courses at Stanford School of Medicine and Icahn School of Medicine at Mount Sinai.^15,16

Next to medical schools, PGT is increasingly applied in the genomics training of prospective pharmacists, with a special focus on pharmacogenomics. At the Eshelman School of Pharmacy at the University of North Carolina and at the University of Pittsburgh, voluntary genotyping is offered to advanced pharmacy and students in the Doctor of Pharmacy (PharmD) program to increase the understanding and acceptance of pharmacogenomics in the clinical context.^17–19 The classes included up to 145 students and genotyping was conducted by a DTC company. At the University of Florida, a smaller number of third-year PharmD students could undergo PGT of selected pharmacogenomic markers as part of the elective pharmacogenomics class.²⁰ The only published genomics course including PGT for undergraduate science students rather than medical or pharmaceutical students took place at the Brigham Young University.²¹ Here, PGT was offered as part of the Advanced Molecular Biology and Genomics course. The students received their personal genotyping data not during, but after completion of the course, assuming that the students’ motivation would be increased by undergoing genetic testing.

Many of the offered courses in the US rely on genotyping performed by private DTC companies.^{12,15, 17–19,21} In Europe, there is no uniform legislative framework covering DTC genetic testing across the 27 individual member states.²² In Germany, genetic testing is regulated by the Genetic Diagnostics Act (Gendiagnostikgesetz, GenDG)²³, which requires testing to be carried out by medical doctors, following written informed consent; in case of preemptive testing, genetic counseling by certified medical doctors is required. Therefore, the GenDG does not allow for delivery of clinically relevant results within a DTC genetic testing framework.²⁴ However, it is allowed to obtain raw genetic data. Additionally, genetic testing for research purposes is exempt.

The present paper describes the implementation and evaluation of the AYPG course at HPI in Potsdam, Germany. Moreover, the AYPG course is compared to the Genomic Medicine (GM) elective course at the School of Medicine of the Technical University of Munich (TUM), which, to our knowledge, was the first course at a German university to include a PGT component in 2017. The AYPG and GM courses were evaluated using questionnaires to assess the students’ attitudes concerning PGT and its usefulness as part of an educational program.

This section describes the implementation of the AYPG course and the course questionnaires.

2.1 Course Implementation

The elective AYPG course was offered to DH Master’s students at HPI. After a course briefing at the beginning of the summer term 2020, covering relevant information about the course and establishing the voluntary nature of the PGT, 16 students in their second year of Master’s studies participated in the course. The course consisted of four three-hour lectures throughout the semester and a five-day block course in August 2020 (Fig. 1). A second iteration of the course was conducted in summer term 2021.

After the first introductory lecture, including information about risks and challenges of genetic testing as well as ethical aspects, students could choose to participate in the PGT. Additionally, a course information sheet and consent form were provided to ensure informed consent (Supplement 1). A study coordinator and honest broker who is not associated with the course coordinated the genotyping process, including communication with students, obtaining consent, and collection of samples for PGT. The honest broker had no access to PGT information and researchers conducting quality control of the genetic data worked with pseudonymized data. Personal data was processed in compliance with the EU General Data Protection Regulation and the Brandenburg Data Protection Act. The described genomics course with PGT opportunity was approved by the ethics board of the University of Potsdam with application number 17/2021.

Genotyping was conducted using the Illumina Infinium Global Screening Array (GSA v3-MD). Throughout the course, only selected single-nucleotide variants (SNVs) were accessible to the students: specific pharmacogenomic markers, SNVs determining the carrier status of rare diseases with recessive inheritance patterns, and backbone SNVs not suitable for determining the individual risk of rare monogenic diseases and limited to assessing risk of common multifactorial traits or diseases (Supplement 2). PGT information was encrypted and password-protected. Only after successful course completion, students could request the password to their own encrypted data from the honest broker. To obtain additional large genotyping data sets, open-access data from the openSNP database was downloaded using the opensnp-cohort-maker pipeline.^25,26 The final openSNP data set included 509,712 variants in 3,581 individuals. Furthermore, students who chose not to participate in the PGT component of the course or who chose not to analyze their personal genetic data during the course were assigned openSNP data for analysis.

2.2 Course Questionnaire

To evaluate the course and to assess the impact of the course on students’ attitudes towards genetic testing and analyses, we established questionnaires that were completed at four different time points: a pre-course questionnaire (Q1) after the introductory course, an intermediate questionnaire (Q2) before working with one’s own data, an end-of-course questionnaire (Q3) after working with one’s own data, and a retrospective questionnaire (Q4) about three months after completion of the course (Fig. 1, Supplement 3).

The overall questionnaire design was inspired by the questionnaires used to evaluate a course including genome sequencing at the Icahn School of Medicine at Mount Sinai.^16,27 The first iteration of the Q1, Q2, and Q3 questionnaires contained questions about the motivation to participate in the course, the usage of personal genome data, usefulness of array-based genotyping data in clinical practice, usefulness of genome sequencing data in clinical practice compared to array-based genotyping data, and opinions on genetic analysis legislation, specifically the German GenDG. A Human-centered design thinking expert from the HPI was consulted to help with the design and implementation of the questionnaires as well as usability testing. After the first three questionnaires were conducted at HPI, the retrospective questionnaire Q4 was shortened considerably, in coordination with co-authors from TUM, to make the questionnaire more accessible and increase participation rates at HPI and TUM. All questionnaires conducted after the summer term 2020 were adapted to this version. To foster truthful answers, the questionnaire invitations and introductions clearly stated that participation is completely voluntary and anonymous. At HPI, pseudonymization was implemented with the help of the honest broker, enabling the anonymous connection of single students’ responses across questionnaires. Students agreed to the complete anonymization of their data after the last questionnaire by deleting the mapping of names to pseudonyms.

The questionnaires were created with SoSci Survey²⁸ and were conducted with students of the AYPG course in the summer term 2020, the TUM GM course in the winter term 2020/21, and the AYPG course in the summer term 2021. Additionally, a slightly adapted version of Q4 was conducted at TUM for all past GM classes (Q4’). HPI questionnaires were conducted in English, which is the teaching language of the DH curriculum, TUM questionnaires in German.

2.3 Statistical analyses

Questionnaire answers were exported from SoSci Survey in CSV format and analyzed using R 4.0.3. To assess significant differences between nominal values in small independent student groups in the short questionnaire versions, Fisher’s exact test was used via fisher.test (base R). To generate effect sizes, Cramer’s V statistic, lsr::cramersV, was computed (lsr 0.5). Effect sizes were interpreted according to the standard interpretation.²⁹ Pseudonymized answers that could be followed across questionnaires at HPI were assessed with Wilcoxon’s Signed Rank test using wilcox.test (paired = TRUE, base R). The test requires ordinal data, therefore, an answer indicating agreement was mapped to a value of 1, disagreement to -1, and no answer or fallback options to NA. Responses on a 5-point Likert scale were summarized accordingly, while a neutral answer was mapped to 0. The r value was used as the effect size, calculated with rstatix::wilcox_effsize (rstatix 0.6.0). The code is available on GitHub (https://github.com/tamslo/sosci-questionnaire-evaluation).

This section compares the AYPG and GM courses and presents the main questionnaire results.

In total, 13 questionnaires were conducted: Q1 to Q4 for HPI 2020, HPI 2021, and TUM 2020/21; as well as the retrospective TUM questionnaire Q4’ (Supplement 4). Most questionnaire results in the following sections focus on the retrospective questionnaires (HPI 2020: N = 10; TUM 2020/21: N = 6; HPI 2021: N = 9; TUM Q4’: N = 47).

3.1 Student populations influenced course content and interests

The GM course was first implemented at TUM in 2017. In particular the implementation of the PGT component, but also parts of the content and structure of this course inspired the AYPG course. However, the curriculum of each course was optimized to the different student populations and learning objectives (Table 1).

Table 1

Comparison of existing genomics courses in Germany.
Course	Analyze Your Personal Genome (Hasso Plattner Institute, HPI)	Genomic Medicine (Technical University of Munich, TUM)
Student population	Second year Digital Health Master’s students	Medicine and Master’s/PhD students of other life and health sciences (different years; see Table 4, question 2)
Course size	16	12–25
Genetic testing strategy	Voluntary genotyping at the Life & Brain research Center, University Hospital Bonn, using the Illumina Infinium Global Screening Array (GSA) v3-MD	Voluntary genotyping (blood withdrawal) at the Institute of Human Genetics (TUM), molecular genetics laboratory, using the Illumina Infinium Global Screening Array (GSA) 24 v1.0
Course design	Four theoretical sessions à 3 hours within two months, followed by a five-day block course with alternating theoretical and practice sessions including discussion rounds. Only a small fraction of personal genetic data was provided during the block course, raw data upon request after the course.	3–4 full days block course with alternating theoretical and practical sessions including discussion rounds. Only a small fraction of personal genetic data was provided during the block course, raw data upon request after the course.
Course content	• Fundamentals of human genetics and molecular biology • Clinical value and ethical implications of genetic testing • Computational analysis of large-scale and individual genotyping data	• Gain knowledge on interpreting and analyzing genetic variants based on public databases • Gain knowledge to understand direct-to-consumer genetic testing options and the associated benefits and risks • Fundamentals of human genetics and molecular biology • Clinical value and ethical implications of genetic testing • Direct-to-consumer-testing products • Ancestry analyses • Interpretation of genetic variants
Pedagogical approaches	• Frontal education • Computational exercises using the openSNP dataset to determine genetic ancestry and calculate a simple genome-wide association study and polygenic risk scores • Small individual projects and presentation on genetic variants and disorders • Panel discussion • Multiple-choice exam	• Frontal education • Computational exercises to determine genetic ancestry using the openSNP dataset • Guided analyses of genetic variants • Student presentations on individual genetic variants and genetic disorders • Discussion rounds • Role play (simulated genetic counseling and adoption of different roles for discussion rounds)
Distinguishing features	Focus on computational analysis of array data using PLINK and additional resources.	Focus on direct-to-consumer testing, interpretation of genetic variants, communication of genetic results to patients, and ethical aspects.

The courses covered introductions to genetics and molecular biology, genetic epidemiology, and statistics, pharmacogenomics, medical genomics, and ethical aspects of genetic testing. The weighting of the individual topics was adjusted according to the students’ prior knowledge. A strong focus of AYPG was computational data analysis using a publicly available data set, covering standard computational tools and analyses in genetics (e.g., genome-wide association analysis and polygenic risk scores). In the GM course, on the other hand, the two major learning objectives were to gain knowledge on interpreting and analyzing genetic variants based on public databases and to understand options for DTC genetic testing and the associated benefits and risks. Accordingly, more emphasis was placed on interpreting DTC test results. Future physicians might be confronted with patients who have purchased DTC products. Thus, a critical understanding of the companies’ products and their benefits and risks is beneficial to medicine students. The analysis of genetic variants of individual patients was discussed in more depth during the TUM course, which is relevant for the clinical setting.

For the AYPG course at HPI, we structured the learning outcomes according to Kraiger et al.³⁰ into cognitive, skill-based, and affective learning (Table 2). While cognitive learning outcomes mainly refer to the accumulation and organization of knowledge, skill-based learning outcomes are related to the acquisition of skills and automaticity, and affective learning outcomes are related to attitude and motivation.

Table 2

AYPG Course schedule containing learning objectives: The course was structured into 12 sessions à three hours. The first four sessions were held as separate lectures during the semester, the last 8 sessions were part of a block course at the end of the semester.
Session Number	CW	Session Topic		Learning Outcomes
Session Number	CW	Session Topic	Cognitive	Skill-based	Affective
1	22	Introduction	• General overview • German Genetic Diagnostics Act (Gendiagnostikgesetz)		• Understand context & decisions facing patients, researchers, and clinicians.
2	25	The Human Genome	• Basics concepts of molecular genetics / the human genome • Genetic diversity & genetic variation • Population structure
3	26	Pharmacogenomics	• Basic concepts and examples of pharmacogenomics	• CPIC guideline interpretation
4	27	Genomic medicine	• Effects of genetic variations, clinical relevance
5–6	32	Genotype-phenotype relationship	• GWAS and GWAS meta-analyses: Basic concepts, examples	• Hands on exercise using PLINK • Explore GWAS summary results using FUMA and MAGMA
7	32	Technical Aspects	• Genotyping & genome sequencing techniques
8	32	Genomic Research, Annotation	• Overview genomic databases / tools and data resources used in annotating and interpreting a personal genome	• Hands on exercise: Exploration and interpretation of selected association results
9	32	Direct to consumer testing	• Basic concepts, Polygenic risk scores		• Understand ethical, legal and social implications of genetic testing
9	32	Disease Risk	• Interpretation of SNV findings (effect size, causal vs. not, multi-allelic models) and the limitations thereof
9–10	32	Polygenic traits and diseases	• Basic concepts, Polygenic risk scores	• Calculate polygenic risk score using the openSNP dataset
10	32	Ancestry	• Population structure in genomes and basic concepts (MDS)	• Infer genetic ancestry of a genome from the openSNP dataset
11	32	Genomic Research, Annotation		• Exploration of selected personal genomic data • Analyze found SNVs associated with disease risk in the context of public databases, literature, and other resources	• Discussion and Presentation
12	32	Q&A with Human Geneticists			• Discussion • Review current understanding of how patients respond to genetic testing results emotionally and behaviorally

The HPI course started with an emphasis on cognitive learning outcomes, including the basics of human genetics, pharmacogenomics, and clinical genomics, preceding the practical modules. In the second part, the focus was set on skill-based learning outcomes in combination with knowledge transfer. Using the openSNP dataset, students learned how to read and filter genotyping data, perform quality control and basic genome-wide association studies, and calculate polygenic risk score using the PLINK tool.³¹ Another focus of the course was variant annotation, interpretation, and risk estimation for individuals and costumers of DTC testing. Here, a combination of cognitive and skill-based learning outcomes was applied by assigning small projects and presentation tasks to each student, accompanied by discussion rounds on specific genetic variants, their annotation, and associated risk. Further discussion on the risks and benefits regarding genetic testing and a direct question and answer session with human geneticists were designed to stimulate a more sophisticated attitude and reflection on the topic.

Regarding the students’ motivation to participate, only few differences between the two courses became apparent (Table 3): compared to TUM students, HPI students were significantly more interested in pharmacogenomics (p = 0.002) and showed a tendency for an increased interest in receiving or analyzing their own genomic data (p = 0.07). Overall, the highest motivation to participate in the courses was the general interest in genomics or genomic analyses (92% at HPI, 91% at TUM).

Table 3

Comparison of motivation to participate in the course between HPI and TUM. The p-values and effect sizes relate to the difference between answers provided by HPI and TUM participants. The information was assessed using HPI questionnaire Q4 2020 (n = 10), HPI questionnaire Q1 2021 (n = 16), TUM questionnaire Q4’ (n = 47), and TUM questionnaire Q1 2020/21 (n = 8).
Topic	HPI (n = 26)	TUM (n = 55)	P-value	Effect size (V)
General interest in genomics/genomic analyses	92% (24)	91% (50)	1	0 (small)
To receive/analyze my own genomic data	85% (22)	64% (35)	0.07	0.19 (small to medium)
Interest in pharmacogenomics (e.g., the impact of genomic variants on medication effects)	77% (20)	38% (21)	0.002	0.34 (medium to large)
To gain knowledge about genomics/genomic analyses to apply it in my professional career	69% (18)	62% (29) ^a	0.61	0.05 (small)
Interest in research topics like genome-wide association studies	58% (15)	45% (25)	0.35	0.09 (small)
To learn about tools for variant interpretation and analysis	54% (14)	64% (35)	0.47	0.07 (small)
Interest in ethical issues in the context of genomic analyses	50% (13)	53% (29)	1	0 (small)
Interest in legal foundations of genomic analyses	50% (13)	44% (24)	0.64	0.03 (small)
Interest in ancestry analysis	46% (12)	47% (26)	1	0 (small)
To receive the credit points	42% (11)	38% (3) ^b	1	0 (small)
Interest in commercial genomic testing (“direct-to-consumer testing”)	38% (10)	45% (25)	0.63	0.04 (small)
To better understand the situation of patients when undergoing genomic testing	38% (10)	45% (25)	0.63	0.04 (small)
^a This question was included for TUM only in the retrospective questionnaire (n = 47) and later included again for HPI questionnaires; it is missing in TUM Q1 2021/22
^b Before the summer term 2021/22, students participating in GM did not receive credit points, therefore the question was only included for TUM Q1 2021/22 (n = 8) when the format was changed from a facultative to a compulsory elective course included in the curriculum.

3.2 Students find the genotyping component useful

Retrospectively, most students felt that genotyping in the course was useful for the learning experience (HPI: 89% [17 of 19], TUM: 92% [49 of 53]); 9 students who indicated that genotyping was a useful learning experience did not participate in the PGT component (Table 4). Almost all students would recommend offering PGT again in future courses (HPI: 95% [18 of 19], TUM: 98% [52 of 53]). 64% (34 of 53) of TUM participants expressed the opinion that a similar course should be an obligatory component of the medicine curriculum at universities.

Table 4

Combined results from retrospective questionnaires. The p-values and effect sizes relate to differences between responses by HPI and TUM indiciduals. The information was assessed using HPI questionnaire Q4 2020 (n = 10), HPI questionnaire Q4 2021 (n = 9), TUM questionnaire Q4’ (n = 47), and TUM questionnaire Q4 2020/21 (n = 6).
Question	HPI (N = 19)	TUM (N = 53)	P-value	Effect size (V)
1. In which year did you participate in the course? (ST = summer term, WT = winter term)	ST 2020: 10 ST 2021: 9	ST 2017: 5 WT 2017/18: 12 ST 2018: 15 WT 2018/19: 6 WT 2019/20: 9 WT 2020/21: 6	–	–
2. Did you study medicine? ^f	Yes: 4 No: 15	Yes: 36 No: 11	–	–
3. Did you participate in personal genotyping as part of the course?	Yes: 17 No: 2	Yes: 44 No: 9	0.72	0.04 (small)
4. From today’s perspective, would you have yourself genotyped again? ^{b, a[3=yes]}	Yes: 6 No: 1	Yes: 6 No: 0	1	0.35 (large)
5. From today’s perspective, would you have yourself genotyped? ^{b, a[3=no]}	Yes: 1 No: 1	–	–	–
6. Did you collect the password for your complete genotype data? ^a[3=yes]	Yes: 13 No: 4	Yes: 35 No: 9	1	0 (small)
7. Do you still plan to collect the password for your complete genomic data? ^{c, a[6=no]}	Yes: 3 No: 1	–	–	–
8. Did you conduct analyses with your own data beyond the course? ^{e, a[6=yes]}	Yes: 7 No: 6	Yes: 14 No: 21	0.52	0.08 (small)
Genetic ancestry ^c	Yes: 2 No: 11	Yes: 2 No: 4	0.56	0.07 (small)
Pharmacogenomics ^c	Yes: 2 No: 11	Yes: 4 No: 2	0.05	0.39 (medium to large)
Wellness traits ^c	Yes: 4 No: 9	Yes: 1 No: 5	1	0.02 (small)
Carrier status ^c	Yes: 1 No: 12	Yes: 2 No: 4	0.22	0.17 (small to medium)
Polygenic diseases ^c	Yes: 0 No: 13	Yes: 2 No: 4	0.09	0.32 (medium to large)
No, I just wanted to receive my data ^c	Yes: 6 No: 7	Yes: 2 No: 4	1	0.01 (small)
9. Do you plan to conduct further analyses with your own genomic data? ^{a[6=yes or 7=yes]}	Yes: 14 No: 2	Yes: 22 No: 13	0.1	0.2 (small to medium)
10. Has this course changed your attitude towards personal genotype analysis?	Yes: 15 No: 2 I don’t know: 2	Yes: 25 No: 24 I don’t know: 4	0.02	0.32 (medium to large)
11. How did the course change your attitude? ^a[10=yes]	More positive: 4 More critical: 11	More positive: 7 More critical: 18	1	0 (small)
12. Would you have participated in personal genotyping as part of the course if it had been offered via a non-European private company, such as 23andMe?	Yes: 7 No: 12	Yes: 11 No: 42	0.22	0.13 (small to medium)
13. Do you think that personal genotyping in the course context is useful for the learning experience?	Yes: 17 No: 2	Yes: 49 No: 4	0.65	0 (small)
14. Would you recommend to offer personal genotyping in future courses again?	Yes: 18 No: 1	Yes: 52 No: 1	0.46	0 (small)
15. How important do you consider the treatment of ethical aspects in the AYPG course?	Absolutely necessary: 12 Rather important: 6 Neutral: 1 Rather unimportant: 0 Completely unimportant: 0	Absolutely necessary: 40 Rather important: 8 Neutral: 5 Rather unimportant: 0 Completely unimportant: 0	0.3	0.19 (small to medium)
16. Do you feel adequately trained to analyze and interpret genomic data beyond the course? ^c	Yes: 8 No: 11	Yes: 2 No: 4	1	0 (small)
17. Would you recommend personal genotyping to relatives or friends who have not taken the course?	Yes: 13 No: 6	Yes: 16 No: 37	0.01	0.31 (medium to large)
18. Should personal genotyping be allowed without genetic counseling by a medical professional?	Yes: 4 No: 15	Yes: 16 No: 37	0.56	0.05 (small)
19. Do you think that a similar course should be obligatory in medical studies at universities? ^{d, a[2=yes]}	–	Yes: 34 No: 19	–	–
^{a[previous question=answer]} Whether question was asked depends on the answer to previous questions
^b Question was not asked in Q4’ and HPI Q4 2020
^c Question was not asked in Q4’
^d Question was only asked at TUM
^e Yes/no question in Q4’ was combined with more detailed answers from other questionnaires: if any analysis was selected by a participant, the answer was counted as “Yes”
^f Was not explicitly asked in TUM questionnaires. Was deferred from more detailed question in Q4’ “What did you study or which education did you have when you participated in the course?” – the answer “Medicine” was counted as “yes”, answers “Genetic and Genomic Counseling” (N = 4) and “Other educational background” (N = 7) as “no”. In TUM Q4 2020/21, all students were medical students.

The majority of questionnaire respondents participated in PGT (HPI: 89% [17 of 19], TUM: 83% [44 of 53]). Many students also collected the passwords to access their full genetic data (HPI: 76% [13 of 17], TUM: 66% [35 of 53]) or still planned to collect it at the time of the questionnaire (HPI: 3 of 4). Overall, less than half of the questionnaire participants who collected their password had conducted their own analyses with their full data at the time of Q4 (HPI: 54% [7 of 13], TUM: 40% (14 of 35)). Of all students who did or still planned to collect their passwords, most still intended conduct further analyses (HPI: 88% [14 of 16], TUM: 63% [22 of 35]).

Of all HPI and TUM Q4 2020/21 survey participants who collected their passwords (N = 19), 6 analyzed pharmacogenomic markers (HPI: 2 of 11, TUM: 4 of 6), 5 wellness traits (HPI: 4 of 11, TUM: 1 of 6), 4 genetic ancestry (HPI: 2 of 11, TUM: 2 of 6), 3 carrier statuses (HPI: 1 of 11, TUM: 2 of 6), and 2 their risk for polygenic diseases (HPI: 0 of 11, TUM: 2 of 6). Notably, only 40% (10 of 25) of students felt adequately trained to interpret genomic data outside of the course (HPI: 42% [8 of 19], TUM: 33% [2 of 6]).

Most students stated retrospectively that they would not have participated in the PGT component if it had been conducted by a non-European private company such as 23andMe: 63% (12 of 19) of HPI students and 79% (42 of 53) of TUM students. Finally, the majority of HPI and TUM participants felt that personal genotyping should not be allowed without prior genetic counseling (HPI: 79% [15 of 19], TUM: 70% [37 of 53]). Interestingly, the two groups of students had different opinions about whether they would recommend PGT to relatives and friends who have not taken the course: 30% (16 of 53) of TUM students would recommend it, compared to 68% (13 of 19) of HPI students (p = 0.01).

3.3 Students changed their attitude towards genotyping throughout the course

The course changed the attitudes of many – especially HPI – students towards personal genotype analyses (HPI: 79% [15 of 19], TUM: 47% [25 of 53]; p = 0.02), with students becoming rather “more critical” (HPI: 73% [11 of 15], TUM: 72% [18 of 25]). This can be confirmed by following the responses across the questionnaires (Supplement 5). HPI students wondered about the interpretability and usefulness of genetic data, specifically whether PGT results are predictive (time point Q1: 2 of 8, Q3: 6 of 8, p = 0.05), whether PGT is useful to inform family members about health risks (time point Q2: 1 of 8, Q3: 5 of 8, p = 0.05), and whether PGT is an opportunity to obtain information that can help them improve their health and wellbeing (time point Q2: 0 of 8, Q3: 3 of 8, p = 0.05). In addition, HPI students felt that genomic counseling should be mandatory for genetic testing (time point Q2: 3 of 8, Q3: 7 of 8, p = 0.07) and that they could learn unwanted information from PGT (time point Q1: 0 of 8, Q3: 3 of 8, p = 0.09). Consistent with the fact that fewer TUM students reported an attitude change, their responses did not change as much over the course as they did for HPI students. In both groups, most students had not yet conducted the analyses they had initially planned at the time of Q3 or Q4.

In the USA, PGT was already implemented as an educational method for university students in several, mostly medical and pharmaceutical schools. By contrast, at European universities, the genomics courses for DH students at the HPI Potsdam and for medical students at the TU Munich are, to the best of our knowledge, the only courses that implemented genetic testing with genome-wide coverage at the time when we collected our questionnaire data. The Medical University of Innsbruck has established a PGT course in Genomic Medicine, modelled on the GM course at TUM as part of their Master’s program in Genetic and Genomic Counselling in 2022.

While the course content differed between the courses at HPI and TUM, the implementation of PGT within the context of the courses was conducted in a similar manner. The use of PGT as part of the offered courses lies outside of the legal restriction of the GenDG, as it is only intended for research and educational purposes. At both Universities, ethical boards explicitly approved the courses and genetic testing for students, performed by a research institution without monetary gain. Students preferred this test setup to the possible alternative of conducting DTC testing via non-European private companies. The implementation described here could serve as a blueprint for universities in Europe when establishing a similar concept for genomics courses at their site.

The concept of personal genotyping for educational purposes can be beneficial for genomic courses of different study programs, as described for the AYPG course at HPI and the GM course at TUM. The idea of adding this educational component to a genomics course was motivated by the assumption that the students’ engagement and learning would be increased during the course and that personal reflection would make them more sensitive to genetic testing in general. Overall, the course questionnaires confirmed that students perceived the PGT component as useful and motivating. The vast majority of students would recommend keeping the PGT component as part of the course and advised to add similar courses to the curriculum of other universities. Nevertheless, the benefits must outweigh any risks associated with the genetic testing, and risks must be minimized to justify the implementation of such educational techniques.

In the university context, these risks include anonymity, confidentiality, and coercion.^11,14,32 The course itself is an elective part of the study program, and PGT was no requirement for course participation. An honest broker system made sure that the genetic data and corresponding personal information were strictly separated. The costs of genetic testing were covered by the Life & Brain Research Center (HPI) and the Institute of Human Genetics (TUM) so that students could undergo PGT regardless of their financial situation. Additionally, genetic testing was not conducted through a private company to mitigate the risk of conflicts of interests and of privacy issues including data leakage.

Furthermore, a major risk of genetic testing in general is that unwanted information may be learned. For the two courses in Potsdam and Munich, just a fraction of low-risk personal genetic information was returned to the students during the course. Only after completion of the course, students could request their complete data, minimizing the risk of making an uninformed decision to explore personal genetic information. At different stages of the courses, risks and benefits of genetic testing were discussed, so that continuous learning of the students regarding this topic was assured.

Most students who participated in the PGT component of the course requested or planned to request the password to decrypt their genetic data, but primarily non-disease related analyses had been conducted at the time of Q4 and Q4’. Many students did not conduct subsequent analyses with their data at all. The questionnaire answers also indicate a change in attitude towards genetic testing among HPI and TUM students. The term “more critical”, that was used when asking students about a change in attitude towards genomic testing, can therefore be understood as “more reflective/aware of ethical implications”, since students were nevertheless interested in receiving and analyzing their own data. These answers presumably reflect the intense discussion of ethical aspects throughout the courses (Table 2, Supplement 6).

As shown with the example of the AYPG and GM courses, the content and structure of genomic courses should be tailored to the specific student population. One important aspect is that the desired learning outcomes differed between medical and DH students. While the study program in Medicine requires a stronger focus on patient-doctor interactions and patient-oriented interpretation of results, DH students were more intensely trained in computational analyses of larger genomic datasets.

Questionnaire results showed that the different focus areas were consistent with the students’ motivation to participate in the course. However, the interest of students may have been modified by the preceding contents and lectures provided at HPI and TUM. For instance, the high interest in pharmacogenomics within the HPI student cohort can be explained by a preceding lecture covering basic concepts of pharmacogenomics that all students from the course were required to attend.

There are further limitations related to the course questionnaires: One major limitation is the small sample size, which is why the results were not further stratified by educational background (and for Q4’ by course years). This may have introduced biases. Moreover, the voluntary survey participants could represent a subset of more positive and motivated or more critical students. While the survey results can certainly provide insights and qualitative trends, no clear conclusions can be drawn from the quantitative analyses. Nonetheless, we consider our approach a valuable evaluation of the course concept.

Finally, understanding genomic information is important not only in the context of the DH and Medicine study programs, but also for the general population, especially regarding personalized health services that provide genetic information to their customers. The majority of students from HPI and TUM supported the idea that genetic counseling should be mandatory when personal genotyping is conducted, as is required by law in Germany. This response is to some degree surprising, as the students received their own raw genetic data without further counseling (albeit with an offer of counseling in case of potentially critical findings) and at least the HPI students would recommend PGT to family and friends who did not even take the course.

While communicating genetic information in digital health applications holds great potential to empower patients and advance personalized medicine, the presentation of content needs to be concise and comprehensible. The StatusPlus patient portal of the University hospital Kiel could serve as a positive example and role model for this.³³ To further support users of digital health applications communicating genetic results, contacts to experts such as human geneticists or genetic counselors could be provided through such applications (e.g., as in the MyGeneRank study³⁴) or integrated into the processes surrounding the applications (e.g., as in services offered by Color Health³⁵). However, with increasing interest and usage of genetic testing opportunities in Germany and elsewhere, obligatory genetic counseling would increase costs for health insurances and possibly exceed the manageable workload of human geneticists. Other educational means, such as online courses or well-designed applications to communicate genetic data, should be considered in the future.

Ethics approval and consent to participate

The described genomics course with PGT opportunity was approved by the ethics board of the University of Potsdam with the application number 17/2021. A course information sheet and consent form were provided to ensure informed consent (Supplement 1). A study coordinator and honest broker who is not associated with the course coordinated the genotyping process, including communication with students, obtaining informed consent, and collection of samples for PGT.

Consent for publication

Students’ participation in course questionnaires was voluntary. It was made clear that data will be published anonymously in an aggregated form.

Availability of data and materials

The questionnaire responses are available in aggregated form, with open text responses removed (Supplement 7). Please contact Tamara Slosarek ([email protected]) with further questions regarding data availability.

Competing interests

Drs. Andlauer and Schurmann are salaried employees of Boehringer Ingelheim Pharma and Bayer AG, respectively. All other authors declare no conflict of interest.

Funding

The research leading to these results has received funding from the Horizon 2020 Programme of the European Commission under Grant Agreement No. 826117.

Acknowledgements

We thank Dr. Per Hoffmann and Stefan Herms (Life & Brain Research Center) for their support with genotyping HPI students as a voluntary part of the course. Further, we thank our honest broker Dr. Jasmin Cirillo (HPI) for all her work, and all additional external lecturers of the HPI course (in alphabetical order): Dr. Aniwaa Owusu Obeng (Icahn School of Medicine at Mount Sinai), Prof. Dr. Johannes Zschocke (Medical University Innsbruck), Dr. Noura S. Abul-Husn (Icahn School of Medicine at Mount Sinai), Prof. Dr. Thomas Meitinger (TUM). We also thank Dr. Jonathan Edelman (HPI) for his help in developing the initial course questionnaires.

Author Contributions

Conceptualization: T.S, S.I., C.S.; Data Curation and Formal Analysis: T.S.; Methodology: T.S., S.I., C.S., H.O.H.; Resources: B.S., T.F.M.A.; Visualization: T.S., S.I.; Writing – original draft: S.I., T.S., Writing – review & editing: C.S., H.O.H., T.F.M.A., B.S., E.B.

Green ED, Gunter C, Biesecker LG, et al. Strategic vision for improving human health at The Forefront of Genomics. Nature. 2020;586(October):683–92. doi:10.1038/s41586-020-2817-4.
Blagec K, Koopmann R, Crommentuijn-Van Rhenen M, et al. Implementing pharmacogenomics decision support across seven European countries: The Ubiquitous Pharmacogenomics (U-PGx) project. J Am Med Informatics Assoc. 2018;25(7):893–8. doi:10.1093/jamia/ocy005.
Widén E, Junna N, Ruotsalainen S, et al. How Communicating Polygenic and Clinical Risk for Atherosclerotic Cardiovascular Disease Impacts Health Behavior: an Observational Follow-up Study. Circ Genomic Precis Med. 2022;15(2):e003459. doi:10.1161/CIRCGEN.121.003459.
Manolio TA, Rowley R, Williams MS, et al. Opportunities, resources, and techniques for implementing genomics in clinical care. Lancet. 2019;394(10197):511–20. doi:10.1016/S0140-6736(19)31140-7.
Wray NR, Lin T, Austin J, et al. From Basic Science to Clinical Application of Polygenic Risk Scores: A Primer. JAMA Psychiatry. 2021;78(1):101–9. doi:10.1001/JAMAPSYCHIATRY.2020.3049.
Lewis CM, Vassos E. Polygenic risk scores: From research tools to clinical instruments. Genome Med. 2020;12(1):1–11. doi:10.1186/S13073-020-00742-5/TABLES/2.
Hoxhaj I, Stojanovic J, Boccia S. European citizens’ perspectives on direct-to-consumer genetic testing: an updated systematic review. Eur J Public Health 2020;(October 2014):1–7. doi:10.1093/eurpub/ckz246.
HPI Digital Health Master’s Program. Accessed May 20. 2022. https://hpi.de/en/studies/before-your-studies/degree-programs/master/master-digital-health.html.
Deci EL, Ryan RM. The “what” and “why” of goal pursuits: Human needs and the self-determination of behavior. Psychol Inq. 2000;11(4):227–68. doi:10.1207/S15327965PLI1104_01.
Haspel RL, Arnaout R, Briere L, et al. A call to action training pathology residents in genomics and personalized medicine. Am J Clin Pathol. 2010;133(6):832–4. doi:10.1309/AJCPN6Q1QKCLYKXM.
Salari K, Pizzo PA, Prober CG. Commentary. To genotype or not to genotype? addressing the debate through the development of a genomics and personalized medicine curriculum. Acad Med. 2011;86(8):925–7. doi:10.1097/ACM.0b013e3182223acf.
Walt DR, Kuhlik A, Epstein SK, et al. Lessons learned from the introduction of personalized genotyping into a medical school curriculum. Genet Med. 2011;13(1):63–6. doi:10.1097/GIM.0b013e3181f872ac.
Sanderson SC, Linderman MD, Kasarskis A, et al. Informed decision-making among students analyzing their personal genomes on a whole genome sequencing course: A longitudinal cohort study. Genome Med. 2013;5(12). doi:10.1186/gm518.
Garber KB, Hyland KM, Dasgupta S. Participatory Genomic Testing as an Educational Experience. Trends Genet. 2016;32(6):317–20. doi:10.1016/j.tig.2016.03.008.
Salari K, Karczewski KJ, Hudgins L, Ormond KE. Evidence That Personal Genome Testing Enhances Student Learning in a Course on Genomics and Personalized Medicine. PLoS ONE. 2013;8(7):1–8. doi:10.1371/journal.pone.0068853.
Linderman MD, Sanderson SC, Bashir A, et al. Impacts of incorporating personal genome sequencing into graduate genomics education: A longitudinal study over three course years. BMC Med Genomics. 2018;11(1):1–12. doi:10.1186/s12920-018-0319-0.
Adams SM, Anderson KB, Coons JC, et al. Advancing pharmacogenomics education in the core pharmd curriculum through student personal genomic testing. Am J Pharm Educ. 2016;80(1). doi:10.5688/ajpe8013.
Frick A, Benton CS, Scolaro KL, et al. Transitioning pharmacogenomics into the clinical setting: Training future pharmacists. Front Pharmacol. 2016;7(AUG):241. doi:10.3389/fphar.2016.00241.
Frick A, Benton C, Suzuki O, et al. Implementing Clinical Pharmacogenomics in the Classroom: Student Pharmacist Impressions of an Educational Intervention Including Personal Genotyping. Pharmacy. 2018;6(4):115. doi:10.3390/pharmacy6040115.
Weitzel KW, McDonough CW, Elsey AR, Burkley B, Cavallari LH, Johnson JA. Effects of using personal genotype data on student learning and attitudes in a pharmacogenomics course. Am J Pharm Educ. 2016;80(7). doi:10.5688/ajpe807122.
Weber KS, Jensen JL, Johnson SM. Anticipation of personal genomics data enhances interest and learning environment in genomics and molecular biology undergraduate courses. PLoS ONE. 2015;10(8). doi:10.1371/journal.pone.0133486.
Hoxhaj I, Stojanovic J, Sassano M, Acampora A, Boccia S. A review of the legislation of direct-to-consumer genetic testing in EU member states. Eur J Med Genet. 2020;63(4):103841. doi:10.1016/j.ejmg.2020.103841.
Human Genetic Examination Act (Genetic Diagnosis Act - GenDG) from 24.04.2009. Bundesrat Printed Matter 374/09.
Statement of the German Society of Human Genetics (GfH) on “Direct-to-Consumer” (DTC) Genetic Testing.; 2011. https://gfhev.de/de/veroeffentlichungen/s-2011_12_02_GfH-Stellungnahme_DTC-Gentests.pdf.
Greshake B, Bayer PE, Rausch H, Reda J. openSNP–A Crowdsourced Web Resource for Personal Genomics. Thornton-Wells TA, ed. PLoS One. 2014;9(3):e89204. doi:10.1371/journal.pone.0089204.
OpenSNP cohort maker. Accessed May 20. 2022. https://github.com/onaret/opensnp-cohort-maker.
Sanderson SC, Linderman MD, Zinberg R, et al. How do students react to analyzing their own genomes in a whole-genome sequencing course?: Outcomes of a longitudinal cohort study. Genet Med. 2015;17(11):866–74. doi:10.1038/gim.2014.203.
Leiner D. SoSci Survey (Version 3.1. 06). Published online 2019.
Kim H-Y. Statistical notes for clinical researchers: Chi-squared test and Fisher’s exact test. Restor Dent Endod. 2017;42(2):152. doi:10.5395/RDE.2017.42.2.152.
Kraiger K, Ford JK, Salas E. Application of Cognitive, Skill-Based, and Affective Theories of Learning Outcomes to New Methods of Training Evaluation. J Appl Psychol. 1993;78(2):311–28. doi:10.1037/0021-9010.78.2.311.
Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4(1):7. doi:10.1186/s13742-015-0047-8.
Korf BR. Pushing the envelope in genomics education. Genet Med. 2015;17(11):857–8. doi:10.1038/gim.2015.20.
Statusplus Test Personal Portal. Accessed May 20. 2022. https://www.statusplus.uksh.de/en.
MyGeneRank – A research study by the Scripps Research Translational Institute. Accessed May 20. 2022. https://mygenerank.scripps.edu/.
Color: A complete platform for healthcare delivery. Accessed May 20, 2022. https://www.color.com/.

Competing interest reported. Drs. Andlauer and Schurmann are salaried employees of Boehringer Ingelheim Pharma and Bayer AG, respectively. All other authors declare no conflict of interest.

Implementation and Evaluation of Personal Genetic Testing As Part of Genomics Analysis Courses in German Universities

Status:

Journal Publication

Version 1

Abstract

Purpose

Methods

Results

Conclusion

Figures

1 Introduction

2 Materials And Methods

2.1 Course Implementation

2.2 Course Questionnaire

2.3 Statistical analyses

3 Results

3.1 Student populations influenced course content and interests

3.2 Students find the genotyping component useful

3.3 Students changed their attitude towards genotyping throughout the course

4 Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1