Sexual Health in Spanish People with Intellectual Disability: the Impact of the Lockdown due to COVID-19

Introduction The lockdown due to COVID-19 affected the sexual health of the people with intellectual disabilities by differentially modifying the frequency and characteristics of people’s sexual activity depending on whether or not they lived with a partner during this period. The aim of this study was to analyze the extent to which the sexual behavior of people with intellectual disabilities (with and without a partner) was affected during the lockdown. Methods The sample consisted of 73 people with intellectual disabilities between 21 and 63 years old (M = 39.63; SD = 10.11). The variables analyzed were the physical, social, and technological environment during the lockdown, sexual appetite, sexual behavior, online sexual activity, and sexual abuse. The data were collected between the months of May and June of 2020. Results The lockdown increased the sexual appetite of a third of the sample (38%), especially the youngest participants. Sexual activity focused on autoeroticism and online behavior, particularly sending nude images of oneself (88%) and viewing pornography (83.6%). Rates of sexual abuse during this period were relatively high (6.8%). Conclusions The sexual activity of people with ID was important during the lockdown, and they had to adapt to the circumstances of isolation in a similar way to the general population. Technological improvements in terms of devices and connection quality at home allowed their sexual behavior to be reoriented, opening the door to new risks for the sexual health of people with ID. Policy Implications Cybersex and the increase in sexual abuse due to confinement are aspects that should be included in programs to improve the sexual health of this group.


Introduction
The sexuality of people with intellectual disability has traditionally been a controversial issue (Franco et al., 2012). The medical model that prevailed for a long time was based on a series of myths that justified the annulment of their experience of sexuality (Jiménez & Huete, 2013). In some cases, their sexuality was equated with that of children, with an absence of sexual desire (Borawska-Charko et al., 2017). In other cases, a lack of control over their impulses was assumed, justifying the prohibition of any expression of their sexuality in order to protect themselves and others (Rodríguez et al., 2006). The overprotective and paternalistic perspective has been replaced by the current models focusing more on providing support on those aspects where is needed rather than imposing prohibitions (Arnau, 2019). There has been an increase in the number of studies designed to determine their demands and needs in order to create programs that provide the necessary support to allow them to fully and responsibly experience their sexuality. Thus, in recent years, professionals who care for people with intellectual disability in occupational or residential centers have asked to receive training in affective-sexual education. These improvements, which have had an impact on the quality of life of people with ID, have not occurred in a generalized way in this group. Instead, they depend on the degree to which their guardians, whether parents in the family unit or professionals in residential centers, have adopted this normalizing perspective (Hoyos & García, 2017). This is a slow process because, especially in the case of parents, it involves people from previous generations (parents of adults with ID are usually over 60 years old) who have grown up with much more restrictive values as far as sexuality is concerned, even in the general population (Manor-Binyamini & Schreiber-Divon, 2019).
In this context of slow and uncertain progress in terms of the sexuality of people with ID, and recognizing that they only have the sexuality they are allowed to have, it is worth asking to what extent people with ID have been affected by the restrictions on mobility and contact with others adopted by many governments, among others, in response to the terrible COVID-19 pandemic.

Sexuality in People with Intellectual Disability
However, before analyzing the implications of the lockdown, let us briefly review what we know about the sexuality of people with ID. Recent reliable and rigorous studies reveal that people with ID show interest in sexuality from very early ages, with sexual desires and needs similar to those shown by the general population (Borawska-Charko, et al., 2017;Gil-Llario, 2019). A study conducted by Gil-Llario et al. (2018) found nearly all adults with ID are interested in sexuality and 96.4% have had a partner. When they do not have a partner, most people with ID express their desire to have one (Gil-Llario et al., 2018) or to have an intimate relationship (Lafferty et al., 2013;Turner & Crane, 2016).
As far as their sexual activity is concerned, between 73 and 85% of people with ID have engaged in sexual intercourse (Baines et al., 2018;Díaz-Rodríguez et al., 2014;Morell-Mengual et al., 2016), with the percentage being higher in women from 40 to 55 years old (Gil-Llario et al., 2018). Higher percentages have been found for masturbation (76-92%), with this sexual practice being considered the most frequent one in this population group, especially in men (Díaz-Rodríguez et al., 2014;Kijak, 2013).
As we have seen, this is a group with a high level of sexual activity, but it is worth asking to what extent this activity is desired and voluntary or, unfortunately, occurs in a context of coercion. The available data on sexual abuse in people with ID, although inconsistent due to different factors such as the criteria established to determine it, cultural differences, etc., (Gil-Llario et al., 2018;Tomsa et al., 2021), reveal high abuse figures in this group. Moreover, there is some discrepancy between professionals' awareness of the experience of sexual abuse and the reality reported by the people who have experienced it (Giménez-García et al., 2017). When the data are self-reported, we find percentages of around 9% for women and 3% for men Mitra et al., 2016), but these percentages increase considerably when reported by professionals (24-28% in women and 16-30% in men) Haydon et al., 2011;Swango-Wilson, 2009). This prevalence could be even higher, given the difficulty people with ID have in identifying sexual abuse. The lack of self-reporting of sexual abuse could be explained by several factors, such as cognitive limitations, communication problems, difficulties in distinguishing between public and private, embarrassment, and negative reactions from the environment when the event is reported. The latter is not a trivial problem because, unfortunately, most cases of abuse are carried out by people close to them who are responsible for their education and care (Diaz-Rodriguez et al., 2014;Martinello, 2014;Medina-Rico et al., 2018). In this regard, recent research shows deficiencies in affective and sexual education in this group, in terms of both conceptual aspects and skills, which makes it difficult to identify indicators of risk of abuse, thus making people with ID much more vulnerable (Byrne, 2018;Gil-Llario, 2019;Sinclair et al., 2015). According to the WHO (2021), children with ID present a risk of sexual abuse 4.6 times greater than their peers without ID.

Sexual Behavior During Lockdown
But what are the impact of COVID-19 on Sexuality and Disability: Are We Closer or More Isolated? Ask Sigmund Hough (2020) in a recently published editorial. Environmental and psychological adverse conditions affect sexual functioning and sexual behavior. Isolation may have made their already complicated sexual interactions more difficult to a greater extent. However, there is a lack of studies designed to analyze the impact of the pandemic on the sexual health of people with intellectual disability (Lokot & Avakyan, 2020).
The context of confinement has highly likely further limited the sexual rights and freedoms among people with ID, a sexually active but highly vulnerable group with restricted access to sexuality (Lokot & Avakyan, 2020). They are routinely forced to develop their sexuality and relationships in a framework of dependency, with a narrow social and private life and significant restrictions (Frawley & Wilson, 2016) that become much more evident during a lockdown like the one caused by COVID-19. This scenario of home lockdown or confinement in residences has deprived people with ID of contact with people other than cohabitants or caregivers, following strict recommendations for social distancing. In turn, the lockdown has contributed to further reducing the limited privacy people with ID tend to have (Rogers, 2009), which can often be almost none. All of this is combined with the existing conservative attitudes about the sexuality of people with ID, especially those held by immediate family members (Bazzo et al., 2007;Morell-Mengual et al., 2017;Tamas et al., 2019), and prohibitions imposed by overprotective parents, or in the case of residential centers, restrictive rules that limit the control people with ID have over their lives (Murphy & Bantry-White, 2020). This restrictive context may have influenced their sexual interest or desire, as well as the actual possibilities of carrying out sexual practices, not only with a partner, but also alone.
Likewise, it is important to keep in mind that, in the best-case scenario, and when a certain degree of privacy has been guaranteed, the situation of confinement may also have led to changes in the repertoire of sexual behaviors, with other sexual activities, such as those carried out online, becoming more important. Although this is still a relatively unexplored area, a growing number of studies show the use of new technologies by people with ID (Chiner et al., 2017;Ramsten et al., 2020). Along these lines, one of the few studies that has explored access to social networks by this group was carried out in Australia. It reveals that the main uses of social networks include making new friends, maintaining existing friendships, and exploring and expressing sexuality (Darragh et al., 2017).
Another consequence of confinement is the increase in the levels of sexual abuse, an aspect observed in the general population, based on the increase in the number of calls for help to services for abuse victims. The lockdown that people with ID have had to experience for months may have made it more difficult to identify situations of risk of sexual abuse, and it may even have aggravated these situations because many cases of abuse occur in the family context (Gil-Llario, 2019; Hughes et al., 2012).

Current Study
For all these reasons, the aim of this study was to analyze the experience of sexuality of people with ID during the period of confinement due to COVID-19. For this purpose, we analyzed the physical, social, and technological environment where these people were confined, as well as the sexual desire, sexual behavior, practice of online sexual behaviors, and presence of sexual abuse. These aspects were analyzed taking into account variables such as gender, age, whether they had a partner, and the availability of electronic devices.

Participants
The sample consisted of 73 adults with mild intellectual disability (n = 43 men and n = 30 women) between 21 and 63 years old (M = 39.63; SD = 10.11); 16.4% were between 21 and 29, 37% between 30 and 39, 30.1% between 40 and 49, 11% between 50 and 59, and 5.5% were more than 60 years old. All the participants were receiving some type of community health resource, either in an occupational center or in a care center. Regarding their sexual orientation, 83.5% identified themselves as heterosexual (79% men and 90% women), 11% as homosexual (14% men and 6.7% women), and 5.5% as asexual (7% men and 3.3% women). A total of 48.8% of the men and 53.3% of the women reported having a stable partner. Table 1 presents the sociodemographic characteristics by gender in greater detail.

Measures
An ad hoc questionnaire composed of 18 items with different response formats was used to assess different facets of sexuality developed during the confinement period due to COVID-19. The instrument is an adaptation of the one used in a parallel investigation carried out with the general population (Ballester-Arnal et al., 2020). The following areas were evaluated: • Physical, Social, and Technological Environment Where the Lockdown Took Place. This section consists of six items that evaluate whether, during the confinement period, the person was in a residential center or with the family and, in the latter case, the type and number of people with whom he/she lived and whether they are the people with whom he/she usually lives. Information is also collected on the availability of a room of their own where they could engage in sexual activity without being disturbed or interrupted, as well as the possibility of having access to electronic devices at any time of day or night. • Sexual Appetite. The intensity of their sexual appetite during lockdown was evaluated with a single item that collects the subjective rating: "How intense was your sexual appetite or desire during the lockdown?." This question was answered on a Likert-type scale with three response options ranging from "less intense than before" to "more intense than before." • Sexual Behavior. This section consists of four items that collect information about different sexual practices carried out during the lockdown: masturbation and sexual relations. The items have a dichotomous response format "Yes" or "No." If they refer to masturbation, there is an additional item that explores the type of masturbation: traditional or manual, with sex toys or watching pornography. If they refer to having sexual relations, the items explore whether they occurred in a physical context or online. • Other Sexual Contexts Online. The practice of online sexual behaviors during the lockdown was evaluated through five items with a dichotomous response format: "I have looked for information about sex education," "I have seen pornographic images or films," "I have entered dating websites," "I sent nude pictures of myself," and "others." • Sexual Abuse. This section consists of two items that ask whether they have been forced to have sexual relations or whether they have forced someone else to have sexual relations during the lockdown period. The item responses have a dichotomous format "Yes" or "No."

Procedure
On March 14, 2020, the Spanish Government declared a state of alarm to manage the health crisis due to COVID-19 (Decree 463/2020), limiting the freedom of movement of people until May 2nd, at which time outings for walking or physical activity were allowed in different time slots according to the age range in towns with more than 5000 inhabitants. Finally, the state of alarm ended on June 21st. After the lockdown ended, a descriptive study was designed based on the study conducted by Ballester-Arnal et al. (2020). Responses were collected using convenience sampling among users of an occupational center for people with intellectual disabilities located in the province of Cadiz (Spain). To determine the participants' eligibility, the following inclusion criteria were established: (a) being over 18 years of age; (b) meeting the criteria of the Diagnostic and Statistical Manual of Mental Disorders for Mild Intellectual Disability (American Psychiatric Association, 2013); and (c) having sufficient communication skills to answer the different questions. The assessment was conducted by the psychologist of the occupational center in interview format and online via videoconference, between the months of May and June of 2020. The average time per assessment was 40 min, 10 of which were devoted to establishing rapport with the participant. The research had the approval of the Deontological Commission of the University of Valencia, and the ethical principles of the Declaration of Helsinki were followed at all times.

Statistical Analyses
Data analysis was performed with the statistical software IMB SPSS Statistics 24. The environment where the confinement took place, sexual appetite, sexual behavior, and sexual abuse were compared based on gender (male or female), age (between 20 and 35 years or between 36 and 60 years), marital status (with or without a partner), and availability of electronic devices (yes or no). The different categorical variables were compared using a chi-square test, and Cramer's V coefficient was used to calculate the effect size.

Physical, Social, and Technological Environment Where the Lockdown Took Place
As Table 2 shows, most of the individuals evaluated spent the months of confinement with their family (87.7%), compared to 12.3% who were in a residential center, with no statistically significant differences based on gender or age range. In the family environment, the people with whom they mainly lived were the father or mother (71.2%) and the brother or sister (54.8%). Although men and women were generally accompanied by the same people, a higher percentage of men were accompanied by their parents (79.1%), compared to women (60%). It is also noteworthy that care provided by a younger relative (niece or nephew) was only the case for men between 36 and 60 years of age. In addition, almost two-thirds of the participants reported having their own room (64.4%). Finally, 54.8% of the sample, with no gender or age differences, had free access to electronic devices at any time of day or night. Thus, taking into account that the percentages are cumulative, and that the majority had access to more than one device, 97.3% had a cell phone, 61.1% had a computer, and 45.2% had a tablet.

Sexuality and Sexual Behavior by Gender
Regarding the intensity of their sexual appetite (see Table 3), men experienced greater sexual appetite during lockdown than women, although the differences between the two groups were not significant (p = 0.436, V = 0.153). All participants, regardless of gender, reported having masturbated during confinement. The most common type of masturbation performed was traditional or manual masturbation (100% in both men and women), followed by masturbation while viewing pornography (23.3% in men and 20% in women). Overall, women reported significantly higher rates of masturbation with sex toys (p = 0.001), with a moderate effect size (V = 0.396). The online sexual behaviors performed most varied by gender; however, the only behavior that reached statistical significance was sending nude images of oneself, where men reported higher rates than women (100% vs. 75%; p = 0.050; V = 0.384). The percentage of men and women who claimed to have felt forced sexually was roughly equivalent. In contrast, when they recognized having forced themselves on another person, the percentage of men was significantly higher than that of women, and these differences were statistically significant (p = 0.033) with a small effect size (V = 0.265).

Sexuality and Sexual Behavior by Age
The intensity of their sexual appetite during confinement varied according to the age range (see Table 4). Thus, individuals between 20 and 35 years of age reported a greater sexual appetite (63%) than individuals between 35 and 60 years of age (22.7%). These differences were statistically significant (p = 0.003), with a moderate effect size (V = 0.408). In general, and in both age ranges, the most common sexual practices were traditional or manual masturbation, followed by masturbation while viewing pornography, masturbation with sex toys, physical sex, and online sex. However, significant differences were only found in masturbation behavior while viewing pornography, in favor of young people between 20 and 35 years of age (p = 0.002; V = 0.208). Finally, it is worth noting that no young people between 20 and 35 years of age reported feeling sexually forced by another person, compared to 10.9% of those between 36 and 60 years of age, although these differences did not reach statistical significance (p = 0.150, V = 0.208).

Sexuality and Sexual Behavior According to Marital Status
The intensity of sexual appetite was mediated by the marital status (see Table 5). In general, people with a stable partner increased their appetite during lockdown, whereas single people experienced a decline, but these differences were not statistically significant (p = 0.077; V = 0.277). Regardless of their marital status, everyone who was evaluated reported having masturbated during the confinement period. However, people who had a partner had higher percentages of masturbation with sex toys (p = 0.172, V = 0.191) and masturbation while viewing pornographic content (p = 0.044; V = 0.265). In addition, only those in a dating relationship engaged in sex physically (12.5%) or online (9.4%) during lockdown, although only the former type of sexual relation reached statistical significance (p = 0.044; V = 0.265). Online sexual behavior also varied depending on the marital status. People with a partner reported higher percentages of sending nude images of oneself (100% vs. 87.5%; p = 546) and viewing pornographic images or movies (87.5% vs. 77.8%; p = 0.353); and single people reported higher percentages of

Sexuality and Sexual Behavior According to the Availability of Electronic Devices
The intensity of their sexual appetite was slightly lower in those who had electronic devices (42.1%) compared to those who did not have this technology (33%), although these differences did not reach statistical significance (p = 0.673; V = 0.097) (see Table 6). In general, people who had electronic devices reported higher percentages of masturbation with sex toys (17.5%) and masturbation while viewing pornographic content (25%) than those who did not have these devices (9.1% and 18.2%, respectively). Sexual relations were strongly influenced by the availability of electronic devices. Thus, people without devices reported higher percentages of physical sex (p = 0.038; V = 0.265), and people with devices reported higher percentages of online sex (p = 0.247; V = 0.188). In addition, as expected, people with electronic devices engaged in a larger number of online sexual behaviors: sending nude images of oneself (100% vs. 81.3%; p = 0.280), viewing pornographic images or movies (87.9% vs. 80%; p = 0.366), searching for sexual information (30% vs. 12.1%; p = 0.066), and searching for a partner on dating websites (15.3% vs. 9.1%; p = 0.445). Finally, the percentage of people who acknowledged having forced themselves on another person was higher in those who had electronic devices (25%), compared to those who did not have this technology (12.1%), although these differences were not statistically significant (p = 0.235, V = 0.163).

Discussion
The aim of this study was to analyze the differential impact of the COVID-19 lockdown on the sexual behavior of a group of people with ID, based on their age, gender, marital status, and access to electronic devices. This research is especially relevant given the scarcity of studies designed to analyze the impact of the pandemic on the sexual health of this group (Hough, 2020;Lokot & Avakyan, 2020). This aspect may have been affected by stereotypes that persist and can limit the healthy sexual experiences of people with ID more than in the general population (Gil-Llario et al., 2021;Lam et al., 2021). Regarding the physical, social, and technological environment where the confinement took place, the results showed that most of the participants were with their families, mainly with parents and siblings. In any case, almost all the people they lived with were people they were related to or lived with regularly. More than half had their own room and access to electronic devices. These data are quite relevant because it is highly likely that the impact of confinement on the experience of sexuality could be influenced by these variables. However, having one's own room does not necessarily mean having a high level of privacy. In fact, some studies conducted with this population show high levels of supervision by family members and caregivers that limit their independence (Brown & McCann, 2018;Retznik et al., 2021).
The results obtained show that the lockdown influenced sexual desire. The differences were more pronounced depending on the age range, increasing significantly in younger people and decreasing in older people. This finding may seem contradictory, but it is not. In general, young people interact with a larger number of people on a dayto-day basis (Merrells et al., 2019;Wilson et al., 2017), and they have younger parents who, in turn, have more normalizing attitudes about sexuality (Manor-Binyamini & Schreiber-Divon, 2019). Thus, the probability of having a freer sexual experience is significantly higher than in older people. Therefore, it is logical that their sexual appetite increased during the confinement period due to the impossibility of satisfying needs that were probably fulfilled more before confinement. Sexual appetite was also greater in those with a stable partner (with whom they did not live) because the lockdown interfered with the possibility of sharing affection and intimacy like before (Baines et al., 2018;Gil-Llario et al., 2018).
Regarding autoeroticism, our results indicate that masturbation was important for all the individuals evaluated, regardless of gender, age, or other variables. This finding coincides with other studies that conclude that autoeroticism is a common sexual practice of people with ID (Díaz- Rodríguez et al., 2014;Kijak, 2013). In general, women reported a greater use of sex toys for this practice, which was not surprising because most of the sexual accessories available on the market are designed and intended for the female anatomy. Hence, some men only use them with a partner, but not individually (Herbenick et al., 2017). Moreover, our results show higher rates of masturbation while viewing pornography in the youngest participants. Intergenerational changes, along with the greater climate of tolerance towards sexuality in this group, make it easier for young people to have access to and use electronic devices (Darragh et al., 2017). As expected, and as occurred in the general population (Ballester-Arnal et al., 2020), the lockdown could have reduced the possibility of having sexual relations with other people. Only individuals who had a stable partner were able to maintain physical relationships, when the confinement took place in the same physical space, or online relationships through electronic devices when they were confined in different locations.
In addition to autoeroticism, the most frequent sexual behavior, our study also revealed a high percentage of other online sexual behaviors, especially among those who had electronic devices. A large number of studies have found these online behaviors to be a normative element in people with ID (Darragh et al., 2017;Ramsten et al., 2020). In general, online sexual activity can have great benefits when it is carried out to connect with friends or seek information about sexuality. (Chadwick et al., 2017), but it can also have negative consequences in the case of sending nude images, due to the sender's vulnerability, or viewing pornographic films, due to the risk associated with the poor quality of the models offered in many of them (Chiner et al., 2017). Due to its novelty, the phenomenon of online risks for people with ID is still in an exploratory phase, and the knowledge is still limited (Normand  -St-Louis, 2016). The practice of sexting was one of the most frequent online activities performed during confinement, especially among men. Gómez-Puerta and Chiner (2021) indicate that people with ID perform this practice, even though they are not sufficiently prepared and have not received the necessary training to manage the risks arising from this behavior. In addition to the lack of training and information, there is a lack of supervision by parents or guardians, often due to their low level of digital literacy (Goli et al., 2020). Finally, the data obtained show that sexual victimization is a phenomenon that has continued to occur during the confinement period. Thus, 6.7% of the people studied felt sexually forced, and 19.2% acknowledged having forced themselves on another person during this period. Undoubtedly, confinement may have favored the presence of this type of violence because, in most cases, the abuse is perpetrated by a member of the immediate family or social environment (Mitra et al., 2016;Tomsa et al., 2021). The abuse rates obtained during confinement are very similar to those reported in other national and international studies that analyze abuse throughout the life cycle, where prevalence ranges from 6 to 9% (Gil-Llario et al., 2018;Mitra et al., 2016). Unfortunately, our data do not allow us to ascertain the degree of severity or whether it involved insistence or consummated abuse. In any case, these data are alarming. Moreover, the number of abuses found could be much higher, given that only a small percentage of all sexual abuses are reported (Willott et al., 2020), due to either shame or lack of awareness and sexual information (Borawska-Charko et al., 2017;Byrne, 2018). In fact, when abuse is reported by a professional, the percentages are four to five times higher than when it is self-reported by the subject . The results obtained show that this type of victimization is not mediated by age or the availability of electronic devices, as indicated in other similar studies (Gil-Llario et al., 2018). Furthermore, no gender differences were found in the percentage of people who said they felt forced, but there were differences in the percentage who acknowledged having forced another person, with this figure being much higher in men. In this regard, there is a strong consensus in the scientific literature that, when a person with ID perpetrates sexual abuse, the person responsible is almost always male (Koçtürk & Yüksel, 2021;Wissink et al., 2018).
This study also has some limitations that should be taken into consideration. First, it is a self-reported and retrospective study, which may result in recall bias or social desirability bias when talking publicly about an aspect as intimate as sexuality. In addition, the sample was obtained through convenience sampling. Caution should be used when generalizing some results because this type of sampling does not guarantee representativeness.

Conclusions
This is one of the first studies carried out with a sample of people with ID, and it has important practical implications. First, it provides information about the impact of COVID-19 confinement on sexual behavior, an aspect that has been studied very little to date, especially in a group whose sexuality is surrounded by numerous myths and false beliefs (Lam et al., 2021). Second, it is evident that sexuality is an intrinsic aspect of people with ID that has continued to be expressed (Borawska-Charko et al., 2017;Turner & Crane, 2016), even though it was modulated by some factors related to the physical, social, and technological environment where the confinement took place. Third, the results reveal that their sexual activity focused on autoeroticism and online behavior, especially viewing pornography and practicing sexting, showing the risks involved in these types of materials and practices without proper training. Finally, this study highlights, once again, that sexual abuse is associated with the family context (Byrne, 2018;Tomsa et al., 2021), given that this lockdown produced episodes of abuse. One of the main practical applications of the results of this study is to identify the aspects that should be addressed as a priority in the prevention of sexual abuse. Affective-sexual education aimed at this group should focus on knowledge of bodily limits and privacy norms. This information could improve the ability to identify abusive behavior perpetrated by people in their immediate environment for whom they feel affection, such as family members or caregivers.
Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature.

Declarations
Ethics Approval The study was approved by the Ethics Committee of University of Valencia (Valencia, Spain) and was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments.

Informed Consent
Participants were informed about the objectives of the survey, completion times, benefits, and risks, as well as about the anonymity of the responses and the right to stop the survey in any point and for any reasons.

Conflict of Interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.