Consenting for retrospective medical records-based research (MR) and leftover tissue-based research (TR) continues to be controversial. The Utilitarian approach, aiming to safeguard scientific integrity of data, maintains that an implicit consent should be adequate (opt-out approach). The Rights approach that emphasizes autonomy and confidentiality demands an explicit consent (opt-in approach) [7
]. In prospective studies, investigators can readily seek verbal or written consent [16
]. In retrospective studies, however, participants might not be practically contactable and contacting them might cause distress [1
]. Research Ethics Committees review may help maintain confidentially of data by permitting only certain research without patients' consent [17
Clinical research centers in Saudi Arabia have, to a large extent, individually adopted Western and international regulations of clinical research such as the US Title 45 Code of Federal Regulation part 46 [3
], the Declaration of Helsinki [18
], and the guidelines of the Council for International Organizations of Medical Sciences [19
]. These regulations permit the Research Ethics Committees to waive the requirement of informed consent for certain types of retrospective research. Although the Research ethics Committees may be guided by theoretical ethical arguments, they remain uninformed about public views on consenting, as empirical ethics studies have not been conducted in this or similar culture.
This survey was conducted to explore the views of Saudis on consenting for MR and TR in the setting of a tertiary care hospital. The views were examined from three perspectives: personal preference, perception of norm, and perception of current practice. The population surveyed was rather balanced (56% patients, 44% companions, from a variety of outpatient clinics). We found that: 1) there is a considerable diversity among Saudi views regarding consenting for retrospective research which may be related to health status, 2) MR and TR are perceived differently in regard to consenting, and 3) the distribution of perceptions of norm was similar to the distribution of perceptions of current practice but different from that of preferences.
We found a large variation in the perceived norm for consenting for medical records and leftover tissue-based research. Respectively, only 38% and 37% of participants required a consent (general or proposal-specific), the rest either did not require a consent (40% and 49%) or believed that medical records and leftover tissue should be used only for personal medical care (23% and 14%). In comparison, 60% of the Canadian public felt that their permission should be obtained before their health information can be used for research [10
]. Respectively, 65.8% and 29% of Americans required consenting for research on clinically derived and research derived personally identifying samples [13
]. In another American study on a special group with serious medical conditions, the majority of respondents (55.4%) disagreed to the use of their medical records for research purposes without their permission [9
]. Finally, a Swedish study showed that 72% of participants thought that a general consent to use stored samples for new studies is enough provided an approval by a research ethics committee [14
]. In this study, participants who required consenting were divided equally between a one-time general consent choice and a proposal-specific consenting choice. To our knowledge, this is the first study to directly compare these two choices of consenting.
We found that although only about 37% of participants required consent, most participants (> 77%, statements 2-7) do not, in principle, oppose retrospective research. This is consistent with findings from other studies [11
]. About 60% and 85% of New Zealanders were willing (answered yes or maybe) to share their identifiable and non-identifiable health information, respectively, with health researchers [11
]. Similarly, American respondents' assent rate for storage and future use of left over samples for research was 89.4% and 97.3%, using a simple statement or a more detailed one time general consent format, respectively [12
]. A review article, examining individual preferences for TR consenting options, also found high willingness (79-95%) to provide one-time general consent [20
This study was specifically designed to compare perception of norm, personal preference, and perception of current practice; and to compare views toward MR and TR. The distributions of personal preferences and perceptions of norm were significantly different for TR and of borderline significant difference for MR; although a rather strong correlation was found between the distributions of both MR and TR. At least on face value, this speaks against norm-compliant society or a low inclination to express preferences that are different from perceived norm (social desirability bias). There are two types of norm, injunctive norm (people's perception of what is commonly approved or disapproved within a specific culture) and descriptive norm (people's perception of what is commonly done). Although the questionnaires were designed to explore the perception of injunctive norm (the rules that should be followed, regardless of what I personally prefer), it is possible that some respondents confused injunctive norm with descriptive norm. This possibility is supported by the finding that there was no significant difference between perceptions of norm and perceptions of current practice, the fact that "current practice" referred to KFSH&RC, a premier center in the area, and the finding that most of the differences between preferences and perceptions of norm in the entire study population may be contributed by the patients subgroup (who have closer knowledge of "current practice"). The strong correlation between perception of norm and preferences suggests that the disagreement between the two is a matter of degree rather than of direction. It appears that the distribution of choices was "shifted to the left", i.e. stricter, for perceptions of norm compared to preferences, indicating a trusting/satisfied population.
On the other hand, Saudis appeared to be more concerned about MR than TR, with the distribution being relatively shifted to the left for MR. The differences between MR and TR were significant from all three perspectives, despite strong correlations between individual scores. This is rather expected as the information in medical records are more profound, more readily identifying, and more related to patient-comprehended information, than the information that can be obtained from leftover tissues. However, this may indicate less awareness of the potential wealth of information in leftover tissues as a source of DNA, which can identify the tissue source even with fully anonymized samples, is a diary of the future, and relates not only to tissue source but also to his/her family. The importance of privacy and confidentiality are well established in Islamic teachings. For example, Al-Qur'an says, "O ye who believe! enter not houses other than your own, until ye have asked permission and saluted those in them: that is best for you, in order that ye may heed (what is seemly). If ye find none in the house, enter not until permission is given to you: if ye are asked to go back, go back: that makes for greater purity for yourselves: and Allah knows well all that ye do."(Chapter 24, verses 27-28), and prophet Muhammad has said, "if you hear from someone something that he does not wish to be mentioned of him then it is a trust, even if he does not request that it is not mentioned." (Musnad Ahmad, Musnad Al-Kabael 26237)
The distribution of consenting choices was not related to age group, gender, education level, or previous research participation. A British study indicated that the proportion of participants who were not concerned about invasion of privacy varied significantly according to cultural background, ethnicity, and socioeconomic status [7
]. However, consistent with our findings, other studies showed that education had no effect on consenting choices [10
]. It may be expected that sick patients looking for innovative treatment are less strict in consenting requirements. Nevertheless, in one study, the majority of patients with several medical conditions disagreed to use their medical records for research without their permission [9
We found that the apparent health status was significantly associated with perception of norm for MR. The choices of the patients were shifted to the left compared to the choices of companions for both MR and TR, which indicates that studies that include patients only may overestimate the desired strictness' in consenting of the public. Companionship was defined in our study as not having a clinic appointment at the time of the study. Since some of the companions are likely to be also patients, it would be expected that the observed difference in consenting choices according to health status may be an underestimate. It was observed that studies restricted to population with serious medical or genetic conditions may be biased toward more favorable attitudes towards research [9
]. Together with our findings, this suggests that although patients may have stricter consenting requirements, the seriously ill who are looking for breakthroughs may be less strict.
The limits of our study include being based on a single hospital and on perception and preferences rather than on actual consenting/participation. Also, the study did not compare identifiable and anonymized data. Current regulations do not require consenting for the use of existing anonymised data in research (it is not considered human subjects research or is exempted from review). Objection to the use of anonymised data in research is expected to be related to objection to research in general or to the purpose of particular research, whereas objection to the use of identifiable data could be related to concerns about privacy and exposure to the risk of harm, in addition. Our study was designed to explore consenting views on identifiable data; these views may be less strict for anonymised data.