Case Summary
An anonymous contactable complainant who described themselves as a concerned UK health professional complained about a named Novartis employee using LinkedIn to promote medicines. The medicines at issue were Entresto (sacubitril/ valsartan) used in adults with chronic heart failure, and Cosentyx (secukinumab) used in adults with moderate to severe plaque psoriasis.
The complainant noted that he/she had received information on LinkedIn that had been shared (‘liked’ posts) and considered it inappropriate for company employees to use LinkedIn to promote information, including studies, about their companies’ products. It was promotion to nonprescribers and the complainant doubted that the materials shared had undergone appropriate internal review. The material at issue included: an article which discussed the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch; Vivinda TV which appeared to be a resource solely intended for health professionals but was being advertised to the public; and advertising of a Novartis press release regarding Phase III data for Cosentyx.
The detailed response from Novartis is given below.
The Panel noted that the complaint concerned LinkedIn activity on an employee’s personal LinkedIn account. In the Panel’s view, it was of course not unacceptable for company employees to use personal LinkedIn accounts and the Code would not automatically apply to all activity on a personal account; if activity was found to be within the scope of the Code, the company would be held responsible.
The Panel noted that material could be disseminated or highlighted by an individual on LinkedIn in a number of ways including posting, sharing, commenting or liking. The algorithms applied by LinkedIn were relevant including whether an individual could opt out of material being disseminated by such algorithms. In the Panel’s view, activity conducted on social media that could potentially alert one’s connections to the activity might be considered proactive dissemination of material. In addition an individual’s activity and associated content might appear in the individual’s list of activities on his/her LinkedIn profile page which was visible to his/her connections; an individual’s profile page was also potentially visible to others outside his/her network depending on the individual’s security settings.
The Panel noted that the complaint concerned three specific activities:
1 Sharing an article on the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch
The Panel noted Novartis’ submission that the hyperlink provided by the complainant had been proactively shared by the named employee with his/ her connections on the LinkedIn platform.
The Panel noted that the original article appeared to have been authored and posted by a journalist at LinkedIn as part of a weekly ‘Premium report’ which highlighted healthcare news. The original post contained a video and written report of an interview with a named Novartis senior leader. The Panel noted Novartis’ submission that the content of the article was focussed on the business approach to a product launch and lessons learnt from the launch of Entresto. The Panel noted that Entresto was mentioned several times, predominantly in relation to the US health environment, its sales and cost-effectiveness data. The Panel disagreed with Novartis’ submission that the article did not position Entresto positively. The article referred to the Institute for Clinical and Economic Review categorising Entresto as cost-effective and the American College of Cardiology and the American Heart Association issuing guidelines which referred to Entresto as the standard of care for certain heartfailure patients, a decision, that the article stated, was often seen as a gold standard in pharmaceutical commercialisation. In the Panel’s view, an employee of Novartis proactively sharing the article with his/ her connections on LinkedIn was considered to be promotion of Entresto, a prescription only medicine, and the ‘share’ and its associated content should have been certified. A breach was ruled.
The Panel considered that on the balance of probabilities not all the employee’s connections would have been health professionals and therefore sharing of the article with his/her network constituted promotion of a prescription only medicine to the public and a breach was ruled. Furthermore, and on balance, the Panel considered that the positive statements in the article could on the balance of probabilities have encouraged members of the public to ask their health professional to prescribe Entresto and therefore a breach was ruled.
2 Advertising VivindaTV, a resource intended for health professionals, to the general public
The Panel noted that the hyperlink provided by the complainant led to a Novartis post on LinkedIn that referred to the latest research in dermatology and referred to Vivinda TV and sessions from a dermatology congress. Readers were invited to register. The linked registration page clearly stated that to create an account the individual had to declare that he/she was a health professional and their country of practice. The Panel noted that neither the original post nor the linked registration page directly or indirectly referred to specific medicines. The Panel noted Novartis’ submission that it appeared the employee in question had ‘liked’ not ‘shared’ this Novartis post. Regardless of whether it was liked or shared, the Panel considered that neither the post nor the linked registration page contained any product related information. The viewer would have to register as a health professional to see further material. The Panel therefore did not consider that the employee’s endorsement constituted promotion of a prescription only medicine to the public; no information about medicines was supplied to the public and no breaches were ruled accordingly.
The Panel noted that the complainant alleged that materials shared had not undergone internal review. The complainant referred to ‘shared (liked posts)’ and thus the Panel considered that the allegation covered both ‘shared’ and ‘liked’ posts. The Panel noted its comments above that the original post and the linked registration page made no direct or indirect reference to a specific medicine and therefore was not considered as promotional material that required certification. Nor did the Panel consider that it was non-promotional material which required certification. No breaches of the Code were ruled.
3 Advertising a Novartis press release about Cosentyx Phase 3 data
The Panel noted that the third hyperlink provided by the complainant led to a Novartis post announcing data in psoriasis at a 2018 dermatology annual meeting. The post itself did not contain any reference to a product but according to Novartis the ‘find out more’ link led to a press release about Cosentyx data from the congress. The complainant referred to a press release and Phase III data for Cosentyx. The Panel noted that Novartis had not provided a copy of the press release. The company submitted that the press release was initiated by the Swiss based headquarters and had not been examined by the UK company.
The Panel noted Novartis’ submission that the post was ‘liked’ by the named employee and not ‘shared’. The Panel noted its comments above about the number of ways an individual could endorse a post which included ‘liking’. The Panel noted Novartis’ submission that following receipt of the complaint it had identified information about how activity on LinkedIn was visible to one’s connections on their feed. Although it appeared that Novartis had known that a ‘share’ would alert an individual’s connections to the activity, it had not realised that a ‘like’ could also alert one’s connections. Novartis submitted that LinkedIn appeared to have an algorithm which decided which ‘likes’ it would alert one’s connections to. The Panel was surprised that this issue had not come to light previously. It was not inconceivable that similar issues might have occurred previously wherein a ‘liked’ post had been disseminated to a Novartis employee’s connection(s). The Panel understood that if an individual ‘liked’ a post it increased the likelihood that the post would appear in his/her connections’ LinkedIn feeds, appearing as ‘[name] likes this’. In the Panel’s view, companies should remain vigilant and needed to ensure that they took reasonable steps to highlight the potential compliance issues that might arise from ‘liking’ certain posts if such posts could thereby potentially be pushed to their connections’ feed.
The Panel noted Novatis’ explanation that the nature of such an algorithm meant that an individual could not anticipate the outcome of ‘liking’ a post therefore Novartis did not accept that ‘liking’ a LinkedIn post was proactively disseminating information in the same way that ‘sharing’ a post was. Novartis acknowledged that if the named employee had ‘shared’ the post it might constitute promotion to the public.
The Panel noted that whilst the complainant had provided a copy of the original Novartis post, beneath which was a list of likes, he/she referred to ‘receiving information’ on LinkedIn. The Panel therefore considered that on the balance of probabilities the employees ‘like’ had been disseminated by the algorithm to his/her contacts and further considered that such dissemination was the subject of complaint.
In the Panel’s view that an algorithm had disseminated an individual’s ‘like’ did not absolve Novartis from responsibility. The Panel considered that the proactive dissemination of a press release about a prescription only medicine to those who were not health professionals or other relevant decision makers promoted that medicine to the public and might encourage such recipients to ask their doctor to prescribe it. Breaches were ruled. The Panel considered that the ‘like’ of the post and its associated content would constitute promotional material and would require certification under the Code. A breach was ruled.
The Panel was mindful of the complex issues that had to be addressed by companies when advising staff about personal social media use. The increasing use of social media, both in personal and business capacity presented challenges. In addition, many social media platforms used algorithms and had settings which individuals and companies might not be fully aware of.
The Panel was aware that the types of activity performed by the named employee on LinkedIn was not uncommon across the industry. In the Panel’s view, employees might feel inclined to endorse articles related to their senior colleagues on LinkedIn or their company’s corporate social media posts and depending on the content such activity might or might not fall within the scope of the Code, therefore companies needed to issue specific and unambiguous guidance on personal use of social media. This was particularly important if UK employees were likely to follow the social media accounts of overseas affiliates which might have codes, laws and regulations that differed to the UK. In the Panel’s view it was very important that companies regularly reviewed such guidance.
In the Panel’s view, the global social media guidance issued by Novartis to its employees prior to the complaint and dated 2016 was open to interpretation. The Panel was concerned that at the time of the LinkedIn activity in question there was no UK local guidance. The Panel noted that after Novartis was notified of this complaint a UK wide communication was sent. The Panel was concerned about the absence of UK specific guidance at the relevant time. The Panel noted its comments and rulings of breaches of the Code as set out above. Overall, the Panel considered that high standards had not been maintained and ruled accordingly. On balance the Panel did not consider that the circumstances warranted a breach of Clause 2.
CASE AUTH/3038/4/18 HEALTH PROFESSIONAL v NOVARTIS
Conduct of an employee on LinkedIn
An anonymous contactable complainant who described themselves as a concerned UK health professional complained about a named Novartis employee using LinkedIn to promote medicines. The medicines at issue were Entresto (sacubitril/ valsartan) used in adults with chronic heart failure, and Cosentyx (secukinumab) used in adults with moderate to severe plaque psoriasis.
The complainant noted that he/she had received information on LinkedIn that had been shared (‘liked’ posts) and considered it inappropriate for company employees to use LinkedIn to promote information, including studies, about their companies’ products. It was promotion to nonprescribers and the complainant doubted that the materials shared had undergone appropriate internal review. The material at issue included: an article which discussed the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch; Vivinda TV which appeared to be a resource solely intended for health professionals but was being advertised to the public; and advertising of a Novartis press release regarding Phase III data for Cosentyx.
The detailed response from Novartis is given below.
The Panel noted that the complaint concerned LinkedIn activity on an employee’s personal LinkedIn account. In the Panel’s view, it was of course not unacceptable for company employees to use personal LinkedIn accounts and the Code would not automatically apply to all activity on a personal account; if activity was found to be within the scope of the Code, the company would be held responsible.
The Panel noted that material could be disseminated or highlighted by an individual on LinkedIn in a number of ways including posting, sharing, commenting or liking. The algorithms applied by LinkedIn were relevant including whether an individual could opt out of material being disseminated by such algorithms. In the Panel’s view, activity conducted on social media that could potentially alert one’s connections to the activity might be considered proactive dissemination of material. In addition an individual’s activity and associated content might appear in the individual’s list of activities on his/her LinkedIn profile page which was visible to his/her connections; an individual’s profile page was also potentially visible to others outside his/her network depending on the individual’s security settings.
The Panel noted that the complaint concerned three specific activities:
1 Sharing an article on the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch
The Panel noted Novartis’ submission that the hyperlink provided by the complainant had been proactively shared by the named employee with his/ her connections on the LinkedIn platform.
The Panel noted that the original article appeared to have been authored and posted by a journalist at LinkedIn as part of a weekly ‘Premium report’ which highlighted healthcare news. The original post contained a video and written report of an interview with a named Novartis senior leader. The Panel noted Novartis’ submission that the content of the article was focussed on the business approach to a product launch and lessons learnt from the launch of Entresto. The Panel noted that Entresto was mentioned several times, predominantly in relation to the US health environment, its sales and cost-effectiveness data. The Panel disagreed with Novartis’ submission that the article did not position Entresto positively. The article referred to the Institute for Clinical and Economic Review categorising Entresto as cost-effective and the American College of Cardiology and the American Heart Association issuing guidelines which referred to Entresto as the standard of care for certain heartfailure patients, a decision, that the article stated, was often seen as a gold standard in pharmaceutical commercialisation. In the Panel’s view, an employee of Novartis proactively sharing the article with his/ her connections on LinkedIn was considered to be promotion of Entresto, a prescription only medicine, and the ‘share’ and its associated content should have been certified. A breach was ruled.
The Panel considered that on the balance of probabilities not all the employee’s connections would have been health professionals and therefore sharing of the article with his/her network constituted promotion of a prescription only medicine to the public and a breach was ruled. Furthermore, and on balance, the Panel considered that the positive statements in the article could on the balance of probabilities have encouraged members of the public to ask their health professional to prescribe Entresto and therefore a breach was ruled.
2 Advertising VivindaTV, a resource intended for health professionals, to the general public
The Panel noted that the hyperlink provided by the complainant led to a Novartis post on LinkedIn that referred to the latest research in dermatology and referred to Vivinda TV and sessions from a dermatology congress. Readers were invited to register. The linked registration page clearly stated that to create an account the individual had to declare that he/she was a health professional and their country of practice. The Panel noted that neither the original post nor the linked registration page directly or indirectly referred to specific medicines. The Panel noted Novartis’ submission that it appeared the employee in question had ‘liked’ not ‘shared’ this Novartis post. Regardless of whether it was liked or shared, the Panel considered that neither the post nor the linked registration page contained any product related information. The viewer would have to register as a health professional to see further material. The Panel therefore did not consider that the employee’s endorsement constituted promotion of a prescription only medicine to the public; no information about medicines was supplied to the public and no breaches were ruled accordingly.
The Panel noted that the complainant alleged that materials shared had not undergone internal review. The complainant referred to ‘shared (liked posts)’ and thus the Panel considered that the allegation covered both ‘shared’ and ‘liked’ posts. The Panel noted its comments above that the original post and the linked registration page made no direct or indirect reference to a specific medicine and therefore was not considered as promotional material that required certification. Nor did the Panel consider that it was non-promotional material which required certification. No breaches of the Code were ruled.
3 Advertising a Novartis press release about Cosentyx Phase 3 data
The Panel noted that the third hyperlink provided by the complainant led to a Novartis post announcing data in psoriasis at a 2018 dermatology annual meeting. The post itself did not contain any reference to a product but according to Novartis the ‘find out more’ link led to a press release about Cosentyx data from the congress. The complainant referred to a press release and Phase III data for Cosentyx. The Panel noted that Novartis had not provided a copy of the press release. The company submitted that the press release was initiated by the Swiss based headquarters and had not been examined by the UK company.
The Panel noted Novartis’ submission that the post was ‘liked’ by the named employee and not ‘shared’. The Panel noted its comments above about the number of ways an individual could endorse a post which included ‘liking’. The Panel noted Novartis’ submission that following receipt of the complaint it had identified information about how activity on LinkedIn was visible to one’s connections on their feed. Although it appeared that Novartis had known that a ‘share’ would alert an individual’s connections to the activity, it had not realised that a ‘like’ could also alert one’s connections. Novartis submitted that LinkedIn appeared to have an algorithm which decided which ‘likes’ it would alert one’s connections to. The Panel was surprised that this issue had not come to light previously. It was not inconceivable that similar issues might have occurred previously wherein a ‘liked’ post had been disseminated to a Novartis employee’s connection(s). The Panel understood that if an individual ‘liked’ a post it increased the likelihood that the post would appear in his/her connections’ LinkedIn feeds, appearing as ‘[name] likes this’. In the Panel’s view, companies should remain vigilant and needed to ensure that they took reasonable steps to highlight the potential compliance issues that might arise from ‘liking’ certain posts if such posts could thereby potentially be pushed to their connections’ feed.
The Panel noted Novatis’ explanation that the nature of such an algorithm meant that an individual could not anticipate the outcome of ‘liking’ a post therefore Novartis did not accept that ‘liking’ a LinkedIn post was proactively disseminating information in the same way that ‘sharing’ a post was. Novartis acknowledged that if the named employee had ‘shared’ the post it might constitute promotion to the public.
The Panel noted that whilst the complainant had provided a copy of the original Novartis post, beneath which was a list of likes, he/she referred to ‘receiving information’ on LinkedIn. The Panel therefore considered that on the balance of probabilities the employees ‘like’ had been disseminated by the algorithm to his/her contacts and further considered that such dissemination was the subject of complaint.
In the Panel’s view that an algorithm had disseminated an individual’s ‘like’ did not absolve Novartis from responsibility. The Panel considered that the proactive dissemination of a press release about a prescription only medicine to those who were not health professionals or other relevant decision makers promoted that medicine to the public and might encourage such recipients to ask their doctor to prescribe it. Breaches were ruled. The Panel considered that the ‘like’ of the post and its associated content would constitute promotional material and would require certification under the Code. A breach was ruled.
The Panel was mindful of the complex issues that had to be addressed by companies when advising staff about personal social media use. The increasing use of social media, both in personal and business capacity presented challenges. In addition, many social media platforms used algorithms and had settings which individuals and companies might not be fully aware of.
The Panel was aware that the types of activity performed by the named employee on LinkedIn was not uncommon across the industry. In the Panel’s view, employees might feel inclined to endorse articles related to their senior colleagues on LinkedIn or their company’s corporate social media posts and depending on the content such activity might or might not fall within the scope of the Code, therefore companies needed to issue specific and unambiguous guidance on personal use of social media. This was particularly important if UK employees were likely to follow the social media accounts of overseas affiliates which might have codes, laws and regulations that differed to the UK. In the Panel’s view it was very important that companies regularly reviewed such guidance.
In the Panel’s view, the global social media guidance issued by Novartis to its employees prior to the complaint and dated 2016 was open to interpretation. The Panel was concerned that at the time of the LinkedIn activity in question there was no UK local guidance. The Panel noted that after Novartis was notified of this complaint a UK wide communication was sent. The Panel was concerned about the absence of UK specific guidance at the relevant time. The Panel noted its comments and rulings of breaches of the Code as set out above. Overall, the Panel considered that high standards had not been maintained and ruled accordingly. On balance the Panel did not consider that the circumstances warranted a breach of Clause 2.
An anonymous contactable complainant who described themselves as a concerned UK health professional complained about a Novartis employee using LinkedIn to promote medicines. The medicines at issue were Entresto (sacubitril/valsartan) used in adults with chronic heart failure, and Cosentyx (secukinumab) used in adults with moderate to severe plaque psoriasis.
COMPLAINT
The complainant noted that he/she had received information on LinkedIn that had been shared (‘liked’ posts) by a named Novartis employee. The complainant did not consider that it was appropriate for company employees to use LinkedIn to promote information including studies about their companies’ own products. The complainant stated that it was promotion to non-prescribers. The complainant doubted that the materials shared had undergone appropriate internal review. The material at issue included: an article which discussed the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch; Vivinda TV which appeared to be a resource solely intended for health professionals but was being advertised to the public; and advertising of a Novartis press release regarding Phase III data for Cosentyx.
When writing to Novartis, the Authority asked it to consider the requirements of Clauses 2, 9.1, 14.1, 14.3, 26.1 and 26.2 of the Code.
RESPONSE
Novartis submitted that social medial platforms were an important channel of communication, through which the industry could and should engage society in high level topics about science and build the reputation of the industry as an essential component of health care. The company recognised that social media played an increasingly important role in the professional and personal lives of its employees. However, the company took its responsibilities under the Code very seriously and, as such, appreciated this and other opportunities to understand the perspectives of others, to re-examine how it conducted its business, and to continually learn and adapt how it might appropriately manage the use of these rapidly evolving platforms and technologies.
Novartis stated that it was very concerned to receive the complaint. Novartis stated that it was clear that whilst company employees had the right to use personal accounts on social media platforms to communicate their own views and perspectives, the Code might, in some circumstances, apply to such posts and that the company might therefore be held responsible for them; each case would need to be decided by a consideration of all of the circumstances. That said, in such a complex and nuanced regulatory environment, where the effect of self-regulation was decided on a case-by-case basis after the fact, it might be difficult for employees to decide how to approach their personal social media activity. This potential difficulty was even more evident in fora such as LinkedIn which was a business and employment network associated with an employee’s professional interests.
Novartis understood the complainant was concerned that information had potentially been shared with him/her by a named Novartis employee through the employee having ‘liked’ some LinkedIn posts. Novartis noted that the complainant was concerned about the use of LinkedIn to ‘... promote information …’ and that he/she was concerned that this had included studies about Novartis products. Novartis also noted that the complainant was concerned that this might constitute ‘… promotion to nonprescribers …’ and was doubtful whether the shared materials had been appropriately reviewed internally.
These concerns were referenced specifically to three discrete LinkedIn activities by the employee.
Novartis fully understood that proactive sharing of detailed information about a company medicine by a company employee would likely constitute promotion, regardless of whether the channel of communication was digital, paper or verbal. Novartis also recognized that in determining whether the activities cited by the complainant in this case were appropriately conducted or not, all of the circumstances should be taken into account, as summarized in both the PMCPA’s ‘Guidance About Digital Communication’ and in Case AUTH/2988/10/17.
Novartis noted that this consideration should include the nature of the material disseminated, the audience in receipt of the material, any product references, the company’s role in creating the material posted and whether the posting was directed, encouraged or otherwise acquiesced by the company. Novartis also believed it important in this case to understand the degree to which information was proactively disseminated as opposed to it being available to view. Specifically, the differences between the LinkedIn activities of ‘post’, ‘share’ (or ‘re-post’) and ‘like’ were relevant considerations.
Hyperlink 1
Novartis noted that the complainant provided a hyperlink and stated that it discussed the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch. The hyperlink led to a ‘share’ by the named employee. Novartis submitted that a share effectively meant an individual had proactively shared content from a third party with his/her ‘connections’ on the LinkedIn platform and was shown as a ‘share’ in the individual’s list of activity provided on his/ her LinkedIn profile page. The employee had not commented on the share and he/she was not the original ‘poster’ of the initial content. Novartis noted that the hyperlink provided by the complainant no longer linked to the ‘share’ in contention and the ‘share’ was no longer visible on the list of activity on the named employee’s profile page. This was because Novartis considered it important that the employee remove the post until this case had been ruled upon.
Novartis stated that the original article was a business news article from a series entitled ‘LinkedIn Premium Report, Healthcare’ which was generated and posted by LinkedIn. The introduction to the article stated that the Premium Report series ‘… highlight industry trends, job moves and healthcare openings …’ making it clear that the material was focussed on the conduct of business and employment within healthcare.
The material appeared to be based upon an interview with a named senior leader at Novartis conducted by a LinkedIn journalist. The post consisted of a short video clip and a written report, which were broadly similar in context and content. The original content was provided.
Novartis submitted that the content of the written article was centred on questions from the LinkedIn journalist as to why Entresto did not meet the expectations of finance analysts. In the opening paragraph the journalist provided the following opinion:
‘It’s never made much sense to me why Novartis’s Entresto, a fairly priced heart-failure drug that was the first new medicine in its category in more than a decade, largely failed to excite cardiologists’ and goes on to describe how the medicine ‘… underperformed in its first two years on the market.’
The named Novartis senior leader explained that it was a struggle to get use in the system, despite having data that led to inclusion in guidelines and which demonstrated the level of cost-effectiveness of the product. He/she described two issues (pricing and the need to enable changes in behaviour when clinicians were provided with a new treatment option) that prevented the expected uptake.
Novartis submitted that the article focussed on understanding the business approaches to pharmaceutical product launches and the business lessons Novartis had learned from a recent launch. Although the further discussion referred to Entresto and its clinical and cost-effectiveness data, the content did not position Entresto positively, did not focus on detailed descriptions of Entresto use or clinical data and did not include promotional claims. Novartis thus did not consider that the article encouraged the administration, consumption, prescription, purchase, recommendation, sale, supply or use of Entresto. In that regard, Novartis did not believe that the proactive sharing of this content by the Novartis employee constituted promotion as defined in Clause 1, or advertising to the public as defined in Clause 26.1 (nor did it meet the definition of an advertisement as provided in Section 3.3 of the MHRA ‘Blue Guide’). Novartis also considered that, as such, it did not need certification as required by Clause 14.1.
As the information provided was business focussed, rather than being about the medicines per se and there was no detailed discussion about how diseases were managed or how treatments were clinically used, Novartis believed that the information provided did not fall into any of the three categories of ‘information to the public’ described in the supplementary information to Clause 26.2 (proactive information about a medicine, reference information or reactive information), or that it constituted ‘educational material for the public or patients … relating to diseases or medicines …’ as covered by Clause 14.3 of the Code. As such, Novartis did not believe that the employee’s re-post of this material constituted a breach of Clause 26.2 or that there was any requirement to examine the ‘share’, as would otherwise be required by the supplementary information to Clause 14.3.
Hyperlink 2
The second hyperlink provided by the complainant led to a Novartis post on LinkedIn which advertised the availability of educational content on dermatology, intended for health professionals, which was accessed via a Novartis proprietary platform called Vivinda TV. The Vivinda TV platform was used to enable registered health professionals to ‘attend’ scientific congresses and/or access their content remotely. A screenshot of the content accessed at the address cited by the complainant was provided.
Novartis could find no evidence that this content was proactively ‘shared’ by the employee (the employee was not named on the post as accessed via the link provided by the complainant and this LinkedIn activity did not appear as a ‘share’ on the list of activity which was provided on the employee’s profile page). Novartis submitted that the post was ‘liked’ by the named employee. When it was accessed, 122 ‘likes’ of this post could be seen. The list of those who ‘liked’ the post could be accessed, only if the viewer clicked in the vicinity of the word ‘likes’. When Novartis accessed this the named Novartis employee was towards the end (details provided) of the list of 122 who had ‘liked’ it. Novartis submitted that, in practical terms, this required scrolling through and scanning many names before seeing the name of the employee. Only his/her name was given. A viewer would need to click on the name to see any detail about that individual and, importantly, to be able to see if they were an employee of a company.
With regard to the link to Vivinda TV itself, Novartis stated that if a viewer followed the advertised link marked as ‘Register Now’, he/she would not see any educational content on dermatology. Instead, he/ she would land on a page which required him/her to create an account. On clicking ‘create an account’ the viewer was taken to a registration page and had to provide details including name, contact details, country and healthcare speciality. He/she must then confirm that he/she was a health professional in that country in order to complete registration.
Regardless of whether this post had been ‘liked’ or ‘shared’ by the Novartis employee, Novartis noted that neither the content of the original post nor the pages it linked to, contained any information about medicines, disease or healthcare and as such the pages were not subject to the Code. Furthermore, Novartis submitted that good practice had been demonstrated in that a viewer must register to gain access to any such content and to register, the viewer must actively confirm that he/she was a health professional.
Hyperlink 3
Novartis stated that the third hyperlink provided by the complainant led to a Novartis initiated news article announcing data in psoriasis at the 2018 American Academy of Dermatology Annual Meeting in San Diego. The news post itself did not contain any actual data or refer to a product, but if the ‘find out more’ link was followed, the viewer was taken to a press release about Cosentyx which gave details about Cosentyx data at the congress. The press release was initiated by the Swiss based headquarters of Novartis, in Basel, on 16 February 2018; as it was not released in the UK and not intended for a solely UK audience, Novartis submitted that the press release had not been examined by the UK company.
Novartis could find no evidence that the article was ‘shared’ by the named employee (the employee was not named on the post as accessed via the link provided by the complainant and this LinkedIn activity did not appear as a ‘share’ on the list of activity which was provided on the employee’s profile page). Novartis submitted that this post was ‘liked’ by the named employee, and when this was accessed, there were 440 ‘likes’ of this post. As above, the list of those who ‘liked’ the post could be accessed; the name of the employee referred to by the complainant appeared well down the list of names (details provided). Novartis noted that the screenshot provided showed 439 likes, not 440. When this was accessed for the investigation, there were 440 likes. However, the screenshot taken at the time of Novartis’ investigation did not save properly and would not open, so a second screen shot was made at a later date in order to provide an attachment for this response.
Novartis could find no evidence that the post which referred to psoriasis data release was proactively shared by the employee. It did not appear as a ‘share’ on the list of activity on his/her profile page and the hyperlink provided by the complainant did not show a ‘share’ (in contrast to the LinkedIn premium article, which was clearly shared by the employee and shown as such in the hyperlink provided by the complainant and the list of activity on the employee’s profile page).
Novartis noted that the complainant clearly stated that the material was ‘shared’ with him/her by it being ‘liked’ by the Novartis employee, but Novartis was not able to explain how the complainant had had this ‘shared’ with him/her, given the absence of this as a ‘share’ on the list of activity on the employee’s profile page and the absence of the employee’s name or recognition of him/her sharing the material at the hyperlink provided by the complainant.
During the course of this investigation, Novartis had identified some information which might help the company and the Panel to understand how the complainant had seen content that a contact had ‘liked’ even if it had not been proactively shared. On the LinkedIn help pages, there was some advice provided to members as to how the visibility and impact of social activity was managed on an individual’s ‘feed’ (the ‘feed’ being the content that was presented to an individual when they accessed LinkedIn). This help page stated that an individual’s LinkedIn feed contained information from his/her network, his/her own likes, shares, posts etc and, importantly, ‘… other information that we believe you may be interested in’. The help page explained that ‘LinkedIn’s systems track and analyze social actions such as writing a post or article, liking content, or commenting on another members posts or articles’ and that ‘This data is used by our algorithms to provide content relevant to you in your LinkedIn feed’.
Novartis stated that it did not have access to LinkedIn proprietary algorithms and it found it hard to understand how an industry could regulate for what appeared to be an Artificial Intelligence approach by LinkedIn in managing the content which was seen by another individual. The very nature of such an algorithm meant that the various potential outcomes of the algorithm could not be anticipated by an individual employee of a company when ‘liking’ a third-party post. Whilst this was important in terms of the industry learning about how to use LinkedIn and similar platforms appropriately, the fact remained that Novartis could find no evidence that this post was proactively disseminated by its employee.
Novartis recognized the principle that, had the press release about Cosentyx data been proactively disseminated by the employee on his/her personal LinkedIn account, this might constitute promotion to the public for which the company might be responsible. However, as stated above, Novartis considered that the employee’s ‘like’ was not proactive dissemination and, as such, did not constitute promotion to the public. Novartis did not believe that the employee’s ‘like’ of this content was in breach of Clause 26.1 and as it would not require certification or examination it denied breaches of Clauses 14.1 or 14.3.
Instructions to Novartis staff about the use of social media
Novartis submitted that its employees were all given clear instruction on the business and personal use of social media. The UK employee handbook, given to all Novartis staff when their contracts were issued, contained some of the major policies about employee relations and professional practices. The global social media guidance for both business and personal use was explicitly referred to in the main text of the document and required employees to read the full guidance provided as an appendix within the document. Novartis noted that an employee’s contract required them to read and accept the content of the employee handbook when he/she signed. The guidance within the employee handbook was taken from two global guidance documents. Novartis confirmed that the named employee had completed training on these documents. Summary and additional information
Novartis noted that it was clear from an interview with the named employee that he/she fully understood the global guidance that had been provided and the application of the Code to social media activity. He/she was clear that his/her intent was never to promote, and that he/she believed that the LinkedIn premium article was an interesting business article rather than information about a medicine or disease, and therefore could compliantly share this article. He/she was unsure about sharing the Vivinda TV advertisement and so decided to ‘like’ it rather than ‘share’ it, in the belief that a ‘like’ would not proactively disseminate the content. He/she realized that sharing the news article about product data release would constitute promotion, and so deliberately did not ‘share’ that, but clicked ‘like’ instead, again in the belief that a ‘like’ would not proactively disseminate the information.
On receipt of the complaint, Novartis considered it important to remind employees of their responsibilities about the use of personal social media channels; it was important to do this before the ruling in this case as this was such an important area to manage appropriately. Action taken had included a UK company-wide communication from a Novartis UK senior leader, to reinforce the guidelines and remind employees that information about a medicine should not be shared by employees on any social media platform.
In the spirit of learning from all dialogue and complaints within the self-regulatory system, Novartis convened a cross-functional team with representation from communications, compliance, medical, legal and pharmacovigilance to look at its existing guidelines and training and to consider any lessons that could be learned from this complaint. The work of this team would be further informed by the ruling in this case.
Novartis did not consider that it had failed to maintain high standards or had brought the industry into disrepute given the evidence above about the three cited pieces of LinkedIn activity in contention, the intent and conduct of the employee in his/ her approach to social media and the company’s approach to use this case to both learn from, and adapt to, this rapidly changing technology. The company denied a breach of Clauses 9.1 and 2.
PANEL RULING
The Panel noted the broad definition of promotion as stated in Clause 1.2; it encompassed any activity undertaken by a pharmaceutical company or with its authority which promoted the administration, consumption, prescription, purchase, recommendation, sale, supply or use of its medicines.
The Panel noted that the complaint concerned LinkedIn activity on an employee’s personal LinkedIn account. The Panel noted that LinkedIn was different to some other social media platforms in that it was a business and employment-orientated network and was primarily, although not exclusively, associated with an individual’s professional heritage and current employment and interests. In the pharmaceutical industry, the Panel noted that an individual’s network might, albeit not exclusively, be directly or indirectly associated with the healthcare industry. In the Panel’s view, it was of course not unacceptable for company employees to use personal LinkedIn accounts and the Code would not automatically apply to all activity on a personal account; whether the Code applied would be determined on a case-by-case basis taking into account all the circumstances including: the content, any direct or indirect reference to a product, how the information was disseminated on LinkedIn, the company’s role in relation to the availability of the content and whether such activity was instructed or encouraged by the company. If activity was found to be within the scope of the Code, the company would be held responsible.
The Panel noted that material could be disseminated or highlighted by an individual on LinkedIn in a number of ways, by posting, sharing, commenting or liking. The algorithms applied by LinkedIn were relevant including whether an individual could opt out of material being disseminated by such algorithms. In the Panel’s view, activity conducted on social media that could potentially alert one’s connections to the activity might be considered proactive dissemination of material. In addition an individual’s activity and associated content might appear in the individual’s list of activities on his/her LinkedIn profile page which was visible to his/her connections; an individual’s profile page was also potentially visible to others outside his/her network depending on the individual’s security settings.
The Panel noted that the complaint appeared to be limited to matters that had been shared or liked.
The Panel noted that Clause 26.1 prohibited the promotion of prescription only medicines to the public. Clause 26.2 stated that information about prescription only medicines which was made available either directly or indirectly to the public must be factual, presented in a balanced way, must not raise unfounded hopes of successful treatment and must not encourage members of the public to ask their health professional to prescribe a specific prescription only medicine.
The Panel noted that the complaint concerned three specific activities of a named Novartis employee on LinkedIn which were considered as follows.
1 Sharing an article discussing the prescribing behaviour of cardiologists and how to price new medicines in relation to the Entresto launch
The Panel noted Novartis’ submission that the hyperlink provided by the complainant had been proactively shared by the named employee with his/her connections on the LinkedIn platform. The Panel further noted Novartis’ submission that the employee did not post the original article on LinkedIn and had not commented on it.
The Panel noted that the original article appeared to have been authored and posted by a journalist at LinkedIn as part of a weekly ‘Premium report’ which highlighted healthcare news. The original post contained a video and written report of an interview with a named Novartis senior leader. The Panel noted Novartis’ submission that the content of the article was focussed on the business approach to a product launch and lessons learnt from the launch of Entresto. The Panel noted that Entresto was mentioned several times, predominantly in relation to the US health environment, its sales and cost-effectiveness data. The Panel disagreed with Novartis’ submission that the article did not position Entresto positively. The article referred to the Institute for Clinical and Economic Review categorising Entresto as cost-effective and the American College of Cardiology and the American Heart Association issuing guidelines which referred to Entresto as the standard of care for certain heart-failure patients, a decision, that the article stated, was often looked at as a gold standard in pharmaceutical commercialisation. In the Panel’s view, an employee of Novartis proactively sharing the article with his/ her connections on LinkedIn was considered to be promotion of Entresto, a prescription only medicine, and the ‘share’ and its associated content should have been certified as required by Clause 14.1. A breach was ruled.
The Panel did not know how many connections the named employee had on LinkedIn and if they were all health professionals; the company made no submission in that regard. However, as it was a personal LinkedIn account, the Panel considered that on the balance of probabilities not all the employee’s connections would have been health professionals and therefore sharing of the article with his/her network constituted promotion of a prescription only medicine to the public and a breach of Clause 26.1 was ruled. Furthermore, and on balance, the Panel considered that the positive statements in the article could on the balance of probabilities have encouraged members of the public to ask their health professional to prescribe Entresto and therefore a breach of Clause 26.2 was ruled.
2 Advertising VivindaTV, a resource intended for health professionals, to the general public The Panel noted that the hyperlink provided by the complainant led to a Novartis post on LinkedIn that referred to the latest research in dermatology and referred to Vivinda TV and sessions from a dermatology congress. Readers were invited to register. The linked registration page clearly stated that to create an account the individual had to declare that he/she was a health professional and their country of practice. The Panel noted that neither the original post nor the linked registration page directly or indirectly referred to specific medicines. The Panel noted Novartis’ submission that it appeared the employee in question had ‘liked’ not ‘shared’ this Novartis post. Regardless of whether it was liked or shared, the Panel considered that neither the post nor the linked registration page contained any product related information. The viewer would have to register as a health professional to see further material. The Panel therefore did not consider that the employee’s endorsement of the original Novartis post by liking it constituted promotion of a prescription only medicine to the public, nor was it contrary to the requirements of Clause 26.2 as no information about prescription only medicines was provided to the public in the original post or linked registration page. The Panel ruled no breach of Clauses 26.1 and 26.2 accordingly.
The Panel noted that the complainant alleged that materials shared had not undergone internal review. The complainant referred to ‘shared (liked posts)’ and thus the Panel considered that the allegation covered both ‘shared’ and ‘liked’ postings. The Panel noted its comments above that the original post and the linked registration page made no direct or indirect reference to a specific medicine and therefore was not considered as promotional material that required certification and ruled no breach of Clause 14.1 accordingly. Nor did the Panel consider that it was non-promotional material covered by Clause 14.3; no breach of that Clause was ruled accordingly.
3 Advertising a Novartis press release about Cosentyx Phase 3 data
The Panel noted that the third hyperlink provided by the complainant led to a Novartis posting announcing data in psoriasis at a 2018 dermatology annual meeting. The post itself did not contain any reference to a product but according to Novartis the ‘find out more’ link led to a press release about Cosentyx data from the congress. The complainant referred to a press release and Phase III data for Cosentyx. The Panel noted that Novartis had not provided a copy of the press release. The Panel further noted Novartis’ submission that the press release was initiated by the Swiss based headquarters and had not been examined by the UK company.
The Panel noted Novartis’ submission that the post was ‘liked’ by the named employee and not ‘shared’. The Panel noted its general comments above about the dissemination of material on LinkedIn and that an individual could endorse a post on LinkedIn in a number of ways including ‘sharing’, ‘liking’ or ‘commenting’. The Panel noted Novartis’ submission that following receipt of the complaint it had identified information about how activity on LinkedIn was visible to one’s connections (their network) on their feed. Although it appeared that Novartis had known that a ‘share’ would alert an individual’s connections to the activity, it had not realised that a ‘like’ could also alert one’s connections. Novartis submitted that LinkedIn appeared to have an algorithm which decided which ‘likes’ it would alert one’s connections to. The Panel was surprised that this issue had not come to light previously. It was not inconceivable that similar issues might have occurred previously wherein a ‘liked’ post had been disseminated to a Novartis employee’s connection(s). The Panel understood that if an individual ‘liked’ a post it increased the likelihood that the post would appear in his/her connections’ LinkedIn feeds, appearing as ‘[name] likes this’. In the Panel’s view, companies should remain vigilant and needed to ensure that they took reasonable steps to highlight the potential compliance issues that might arise from ‘liking’ certain posts if such posts could thereby potentially be pushed to their connections’ feed. The Panel noted Novartis’ explanation that the nature of such an algorithm meant that an individual could not anticipate the outcome of ‘liking’ a post and therefore Novartis did not accept that ‘liking’ a LinkedIn post was proactively disseminating information in the same way that ‘sharing’ a post was. Novartis acknowledged that if the named employee had ‘shared’ this post it might constitute promotion to the public.
The Panel noted that whilst the complainant had provided a copy of the original Novartis post, beneath which was a list of likes, he/she referred to ‘receiving information’ on LinkedIn. The Panel therefore considered that on the balance of probabilities the employee’s ‘like’ had been disseminated by the algorithm to his/her contacts and further considered that such dissemination was the subject of complaint.
In the Panel’s view that an algorithm had disseminated an individual’s ‘like’ did not absolve Novartis from responsibility. The Panel considered that the proactive dissemination of a press release about a prescription only medicine to those who were not health professionals or other relevant decision makers promoted that medicine to the public and might encourage such recipients to ask their doctor to prescribe it. Breaches of Clauses 26.1 and 26.2 were ruled.
The Panel considered that the ‘like’ of the post and its associated content would constitute promotional material and would require certification under the Code. A breach of Clause 14.1 was ruled.
The Panel was mindful of the complex issues that had to be addressed by companies when advising staff about personal social media use. The increasing use of social media, both in the personal and business capacity, presented challenges. In addition, many social media platforms used algorithms and had settings which individuals and companies might not be fully aware of. In Case AUTH/2851/6/16, which related to a posting on LinkedIn, the Panel considered that the fact that it had occurred as a result of an algorithm did not absolve the company from responsibility.
The Panel was aware that the types of activity performed by the named employee on LinkedIn was not uncommon across the industry. In the Panel’s view, employees might feel inclined to endorse articles related to their senior colleagues on LinkedIn or their company’s corporate social media posts and depending on the content such activity might or might not fall within the scope of the Code; therefore, companies needed to issue specific and unambiguous guidance on personal use of social media. This was particularly important if UK employees were likely to follow the social media accounts of overseas affiliates which might have Codes, laws and regulations that differed to the UK. In the Panel’s view it was important that companies regularly reviewed such guidance.
In the Panel’s view, the global social media guidance issued by Novartis to its employees prior to the complaint and dated 2016 was open to interpretation; it stated:
‘In general you can share public Novartis posts. But be aware that any comments you add to a Novartis post may create a risk of violation of rules regarding the promotion of our products or our company. You must also understand and follow any divisional or local guidance that may limit your ability to share such posts’.
The Panel was concerned that at the time of the LinkedIn activity in question there was no UK local guidance. The Panel noted that after Novartis was notified of the present complaint a UK company wide communication was sent by a UK senior leader which stated:
‘Please do not post, re-repost or share content that makes any reference to a specific medicine, including Novartis products. This includes product-specific information emanating from Novartis corporate social media feeds.’
The Panel was concerned about the absence of UK specific guidance at the relevant time. The Panel noted its comments and rulings of breaches of the Code as set out above. Overall, the Panel considered that high standards had not been maintained and ruled a breach of Clause 9.1. On balance the Panel did not consider that the circumstances warranted a breach of Clause 2 and ruled accordingly.
Complaint received 25 April 2018
Case completed 5 December 2018
