OPEN DATA Tug of war for antiviral drugs data Julia Belluz hears from both sides in the lengthy battle between the drug giants and researchers that led to the release of full clinical trial data on neuraminidase inhibitors. Does the release of these files herald a more transparent era?

om Jefferson can recite the details of is when I started thinking there is something accessing all the evidence necessary to answer every published study on the neurami- wrong,” Jefferson now recalls. He felt duped and the question of whether neuraminidase inhibi- nidase inhibitors oseltamivir (Tami- worried that his previous work was misleading tors did what the drug giants claimed took four flu) and zanamivir (Relenza): date of physicians, patients, and public health authori- fraught years, dozens of letters and emails to key publication, authors’ names, journal ties. “I don’t like being made a fool,” the former stakeholders within the drug companies and out- Ttitle, research question, methods, funder, and British military physician told the BMJ. “I trusted side, regulatory changes, and public and politi- findings. But what has occupied the mind of literature. I trusted people who were doctors cal pressure through open data campaigns by the physician-academic from Cochrane’s Acute and researchers. I trusted the archive. I trusted the BMJ and the AllTrials advocacy group. “We Respiratory Infections Group for the better part Roche.” were met with refusal,” says Jefferson. The drug of five years is another detail about the science That trust evaporated. In December 2009, companies stalled and stonewalled, releasing behind the influenza antivirals: that much of Jefferson and the Cochrane group published an their data only in incomplete pieces and break- it has been missing or updated review on neu- ing promises about transparency along the way. unpublished. raminidase inhibitors in Suddenly, last year, the tug of war ended. Both Before he knew bet- the BMJ.1 This time, the GSK and Roche gave the Cochrane group what ter, in 2006, Jefferson researchers asked Roche are believed to be complete sets of clinical study waded into the evidence for raw data from the reports and regulator documents on their neu- and released a Cochrane unpublished trials. Roche, raminidase inhibitors with no strings attached. systematic review about in exchange, asked them For the researchers, this seemed like an about the neuraminidase inhibi- to sign a confidentiality face. “After all this worry and concern about tors based on published agreement with a secrecy data protection, they just couriered them over to research findings. He and clause. The academics me,” said Carl Heneghan, director of the Centre his coauthors found that refused and went ahead of Evidence-Based Medicine at Oxford University, the drugs decreased the with their review, noting who was also working on the Cochrane reviews. risk of influenza complica- the inconsistencies in the Early last year, five CDs arrived at his Oxford tions and hospital admis- The drug companies stalled and evidence and omitting the University office from GSK, and by September, sions in adults. stonewalled, releasing their studies based on unpub- six more were delivered from Roche. In total, By 2009, however, data only in incomplete pieces lished trials. Without the the discs contained more than 160 000 pages of Jefferson’s opinion had and breaking promises about raw data about oseltami- clinical trial information about zanamivir and started to shift. Europe vir, they wrote that the drug oseltamivir. These data underpinned the two was in the clutches of the transparency along the way may work no better than new reviews on the drugs, which have much less influenza A/H1N1 panic, Tom Jefferson, Cochrane aspirin. They also could favourable conclusions—such as the new con- and he and his peers were Collaboration not say whether it truly cerns about neuropsychiatric harms—than the asked to update their reduced the risk of serious early studies.2 3 The researchers are grateful, but analysis of the flu drugs. Just as he was stepping complications and hospital admissions. they wonder why the companies finally decided into the research again, a letter arrived from a to relent and reveal unpublished data that had Japanese paediatrician, Keiji Hayashi. He was Data promises been hidden from public view for so long. concerned that he could not verify the data that On 8 December 2009, the same day the review supported some of Jefferson’s conclusions about was released, Roche announced that it would Unavoidable delay oseltamivir. In particular, he pointed out that the give the Cochrane team full clinical study Representatives of the drug companies tell a claims that the drug reduced secondary compli- reports about oseltamivir “within the coming different story. They describe feeling bullied cations and hospital admissions were based on days.” (These documents are usually prepared and misunderstood by the researchers. Instead a 2003 analysis coauthored by Roche employ- by industry for regulators and include far more of stalling, pharmaceutical executives on the ees. Hayashi also noted that eight out of the 10 detailed reporting of trials than appears in pub- oseltamivir and zanamivir files say that they trials included in that analysis were never pub- lished articles.) But the dialogue between Roche were scrambling behind the scenes to figure out lished in peer reviewed journals. In other words, and Cochrane didn’t end with that promise. how to respond to a new kind of request for more Cochrane’s previous studies on the influenza Instead, it marked the beginning of a lengthy granular clinical trial data. They argued that pri- treatments were based on an incomplete and tug of war over data. By mid-2010, GlaxoSmith- vacy and legal concerns, logistical challenges, possibly biased picture of the evidence. Kline entered the fray, promising to hand over and miscommunications between the two camps The Cochrane team didn’t have an answer its unpublished research about their influenza halted progress. The companies didn’t crack for Hayashi. They didn’t even know about the antiviral, zanamivir. under pressure last year; instead they describe a existence of the unpublished research. “This According to the Cochrane researchers, slow and gradual learning curve that culminated

BMJ | 12 APRIL 2014 | VOLUME 348 15 OPEN DATA

Put yourself in the position of a scientist within GSK trying to do this because we have said we’re going to do it—and being treated as though somehow your intent was bad, which is how they felt . . . It’s not easy to motivate a group of people who are being beaten Patrick Vallance, GSK president of pharmaceuticals R&D

in the landmark release of their data trove. “It ing to a log jam of communications.” seems [the data have] been dragged out of the Vallance says that the Cochrane group inter- company with everyone kicking and screaming preted the mistake in another way: “The next not to let it go,” said GSK’s president of phar- thing that we saw was the [comment] in the maceuticals research and development, Patrick BMJ saying they were appalled by the degree Vallance. “That was not my perception at all.” of the redaction. Frankly, as a lifelong academic From Vallance’s vantage point, GSK was researcher, I would have thought that the right keen to give the Cochrane group the data thing to do was to write to us and say, ‘This is they wanted all along—but it first had to find over-redacted.’” it. Doing so was a herculean task. “These are Instead of prompting swift action, the letters studies going back a long way, stored in all sorts had the opposite of the desired effect internally. of different parts, in all sorts of different places. “Put yourself in the position of a scientist People don’t expect to have to go and pull out within GSK trying to do this because we have everything again.” GSK estimates 15 to 20 said we’re going to do it —and we’re absolutely people worked on the file part time over three clear we want to do it—and being treated as years. “Because these workers were doing the though somehow your intent was bad, which tracking down in addition to their day jobs,” is how they felt.” He adds: “It’s not easy to moti- said Vallance, “There may have been an ele- vate a group of people who are being beaten.” ment in sometimes not feeling this was the top These miscommunications slowed things of their priority list.” down, says Vallance. “What was interesting to When GSK found all the information the me was the degree of mutual misunderstanding Cochrane group requested, the drug company as to what was actually required. At one stage, offered the batch of files to the researchers if I saw a very formal email exchange with long as those from global collaborative research they would sign a confidentiality agreement. gaps between emails rather than what seemed initiatives—were typically signed under confi- “Cochrane weren’t willing to sign something to me was the obvious way to get to the real dentiality agreements. that said they’d keep it to themselves,” said issue: to pick up the phone and talk to each There was also the ever present but little Vallance. The academics wanted to be able other and say what is it that you haven’t got mentioned problem of competition among the to publish the data so that others could repli- that you now need.” pharmaceutical giants. Roche did not want to cate their findings. They argued this would be Even without communications mishaps, be seen to be moving out of step with the rest a giant step towards open data about public Roche representatives say that until very of industry, Clinch says. “When it was suffi- health medicine that had been stockpiled to recently the conditions for data sharing weren’t ciently mature, when we were clear about the the tune of billions globally. quite ripe. “Looking back to 2009, bear in mind direction we wanted to take, the direction that For the drug makers, this presented a chal- how different the environment that we oper- society could be comfortable with us taking, we lenge: they now had to figure out how to con- ated in then was compared to how it is now,” started to share the CSRs.” ceal patient information in the clinical study said Barry Clinch, UK clinical lead for oseltami- reports, which are usually written for regula- vir at Roche. Back then, few outside of industry New era of openness tors and not the public. Jeffrey Helterbrand, the and the regulators had even heard of a clinical Though the drug companies and the research- global head of biometrics at Roche, says, “Our study report, and there was no critical mass ers view the road to open data very differently, main conversations were around, ‘How do we regarding open data in the pharmaceutical they agree that the release of the oseltamivir balance protecting patient privacy versus meet- sector. “We are now in a situation where gov- and zanamivir files heralds a more transparent ing the altruistic desires of patients and society ernments, health regulatory agencies, patients, era. The Roche and GSK executives say that the to have greater access to this kind of informa- and the industry at large all have a way of look- experience has helped them shape systems for tion?’” ing at data sharing now that is completely dif- making their science more freely available. As Similarly, Vallance says, “There was an issue ferent to how it was in 2009. At that point, we Helterbrand puts it, “I’m optimistic this is a about—not whether to do it—but how to do it in were used to a very simple relationship with our big step forward, not just for industry but for a way to protect confidentiality.” GSK hired a regulatory bodies.” researchers in general who deal with clinical third party to redact the zanamivir trials. “That Clinch called the Cochrane researchers’ trial data.” had to be done form by form,” he explains, and request for clinical study reports unprece- In 2012, GSK announced a web portal for it was a learning process. At first, he says, the dented. “We didn’t have an equivalent way to sharing anonymised patient level data, which people hired to do the job overinterpreted pri- respond to a request that the Cochrane group inspired the recent launch of the joint web- vacy laws and blacked out more pages than was was giving to us.” He claims that the company site ClinicalStudyDataRequest.com, where necessary. “That fuelled suspicion on the other had, before this, little experience sharing large independent researchers can request patient side that something was deliberately being hid- amounts of patient level information with non- level data from clinical studies by GSK, Roche, den,” he says. “I could sense that that was lead- statutory bodies. Other similar requests—such Boehringer Ingelheim, Novartis, Sanofi, and

16 BMJ | 12 APRIL 2014 | VOLUME 348 OPEN DATA

about the effectiveness of flu drugs. “The key thing is that only two out of 10 studies on Tami- flu were ever published. That’s not to do with individual patient data.” (Roche responded by saying that, unlike today, 10 years ago it was not standard policy within industry or Roche to publish all its clinical trial data.) Over the years Jefferson counted at least 12 reasons from Roche why it couldn’t hand over its data. These include everything from patient confidentiality to the Cochrane group’s failure to sign a confidentiality agreement to Roche’s belief that the researchers had all the data they desired. “This is a smokescreen,” Jefferson says. “How many more stories are we going to hear?” If anything, Jefferson says his involvement Igho Onakpoya, Cochrane researcher, is finally able to sift in the battle for the oseltamivir and zanamivir through the complete data data has made him more sceptical. “I do not trust fellow researchers. I don’t trust journals. I don’t trust the literature. Everything for me is marketing and publicity, unless proven other- wise.” Even with the latest transparency initiatives, industry still controls the data it produces. It decides who gets access, Jefferson says, and

who doesn’t. If clinical study reports are THOMAS MARK released to third party researchers, and com- ViiV Healthcare. (To date, 13 requests for GSK Looking forward, he is optimistic that these panies have varying policies about redactions, data have been approved and there are 16 gestures will lead to warmer relations with aca- vital data points—such as those about harms— making their way through the system. None demia. “I hope as we get to the new era, that might remain hidden. As Doshi wrote in a has been turned down. Roche has received suspicion will break down and will allow a peer recent BMJ analysis, despite the widespread four requests and all are currently being scientist to peer scientist interaction.” practice of off-label prescribing, all companies considered.)­ The academics also acknowledge progress, share trials only of approved indications and Third party researchers can also now request but they say any thawing of relations may take not those examining off-label uses.4 clinical study reports from Roche, while GSK some time. “I would wish for warmer relations These individual failures to share data, says has committed to post documents automati- and less scepticism as well, which will hope- Jefferson, belie systemic information inequities cally alongside their online trial registration. fully occur naturally as trust grows,” says Peter at the core of medicine. Though the battle for GSK also leads a steering committee—including Doshi, assistant professor at the University of data left him exhausted, and “cost my family representatives from academia and other drug Maryland School of Pharmacy and associate and me a lot,” he wouldn’t give up because companies—which is working on getting an editor at the BMJ. “But the track record is that of what the Cochrane reviews of oseltamivir independent, third party such as the Wellcome of fundamental problems in transparency, reveal: a gulf between conclusions based on Trust to govern access to with serious ramifica- syntheses of published research findings and data. I’m optimistic this is a big step tions to patient safety, those based on raw data. And he won’t rest Still, not all drug com- forward, not just for industry but which have led to dra- until that gap is closed. panies are on side with for researchers in general who matic declines in trust.” “Tamiflu involves everyone,” he says. “It the transparency push. deal with clinical trial data For these reasons, and involves public health bodies worldwide. It Last year, AbbVie sued Jeffrey Helterbrand, global head of because the Cochrane involves journals . . . the research community the European Medicines biometrics, Roche group was so geographi- . . . the pharmaceutical industry. All these Agency to block the cally dispersed, Doshi actors are responsible. The system is broken. It agency’s plan to release clinical trials data from says it was important to get all negotiations can’t be fixed. It has to be redesigned.” drug companies. Representatives from GSK and with industry in writing—even if it slowed down Julia Belluz Knight science journalism fellow, Roche would not comment on the case but said the process. Massachusetts Institute of Technology, Cambridge, MA, that they believe the trend among drug compa- Others remain sceptical about the drug USA nies is moving in the opposite direction. Val- companies’ supposed difficulties with redact- [email protected] Competing interests: None declared. lance observed a pack effect among industry ing study reports and establishing methods Provenance and peer review: Commissioned; not externally leaders. Everyone is going to be sharing data, for data sharing. Iain Chalmers, one of the peer reviewed. he said. “The fact we’ve managed to do it shows founders of the Cochrane Collaboration and an References are in the version on bmj.com. it’s possible; it shows it’s not this great risk that advocate of open data, says that the so called Supported by a grant from the Open Society Foundations. people have sometimes claimed it is.” challenges were just diversions from core issues Cite this as: BMJ 2014;348:g2227

BMJ | 12 APRIL 2014 | VOLUME 348 17 ANTIVIRAL DRUGS Tamiflu: “a nice little earner” Is the story of oseltamivir, and similar antiviral drugs, a classic one of big pharma greed? Andrew Jack finds a more nuanced reality KATIE COLLINS/PA KATIE

iewed with hindsight, the story of pandemic preparations. The US alone spent Pandemic sales patterns flu antiviral drugs seems to be a more than $1.3bn buying a strategic reserve Indeed, oseltamivir sits awkwardly in the com- classic story of “big pharma” greed. of antivirals.1 Most have never been used, and pany’s portfolio, and in meetings with analysts On the back of media hype and today the US stockpile consists of more than and investors the company typically presented its unfulfilled fears, a new medicine 65 million treatments. In the UK, the govern- past financial performance and future guidance Vbacked by only modest clinical data became ment spent £424m for a stockpile of about 40 excluding oseltamivir to avoid confusion. The a “blockbuster” generating billions of dollars million doses.2 drug offered lucrative but “lumpy” sales that rap- for its producer. The reality is more nuanced, Not all of Roche’s sales were profits. In its idly surged—requiring a substantial investment and the tale could easily instead have been a peak year of 2009, Roche reported revenues in additional manufacturing capacity—and then commercial and public health nightmare. from osteltamivir of $3.6bn. Like all drug com- fell again. That contrasts with the classic lifecycle There is no doubt that Roche, the Swiss panies, it does not disclose profits on individual of medicines, for which demand typically grows based pharmaceutical group, and its share- products. Yet it paid out more than $50m in steadily until their patents expire while profits holders, have done very well from oseltamivir initial development costs to Gilead, the US bio- rise as their ongoing costs fall. While financial (Tamiflu). The drug became one of the most tech company that discovered the drug. In 2009 analysts’ reports in the late 2000s referred to widely recognised medicines in the world as alone, it transferred royalties to the company oseltamivir sales, few devoted much attention to concern grew about a new flu pandemic in the of $393m. Since oseltamivir’s launch, it has the drug, and there is little evidence that the drug middle of the previous decade. contributed royalties in excess of $2.2bn. had a significant effect on Roche’s share price. As one City financial analyst puts it: “Tamiflu Oseltamivir is also relatively costly and com- Nevertheless, oseltamivir’s margins were sub- was a nice little earner. It reflected opportun- plex to manufacture, with a multiple step syn- stantial by pharmaceutical industry—let alone istic action by a multinational corporation, thesis that begins with extracting raw material broader corporate—benchmarks. Costs were kept which was able to be a little bit sharper in its from the star anise plant. Roche had to take low because there were relatively few expensive marketing practices than you could be now, account of direct marketing and storage costs, as clinical trials. Marketing expenses were also given the debates over the disclosure of clini- well as a share of overheads across the business. limited because so many contracts were negoti- cal data and how effective the drug was.” Under pressure to provide the drug to govern- ated with a handful of decision makers in gov- Yet other related antiviral drugs in the same ments in large volumes as fear grew of a lethal ernments rather than the usual deployment of class of neuraminidase inhibitors failed to take pandemic, Roche sold oseltamivir for stockpiles large numbers of sales representatives proposing off; oseltamivir itself was nearly a flop; public at a discounted price of €15 per adult course in medicines to a multitude of doctors and health- health pressures capped its pricing; and policy higher income countries and €12 in middle and care departments in each country. makers and the drug company itself struggled lower income ones. Unusually, it also sold bulk By 2009—the last time of any substantial dis- with scant clinical information on efficacy. Its pharmaceutical ingredients to governments at closure—96 countries had stockpiled enough success was linked to the unprecedented pur- a discount to the price of the finished tablets in oseltamivir to provide support for an estimated chase of such a large volume of drugs to prepare packets, as a way to prolong the shelf life. 350 million people around the world. Once pres- for a future pandemic threat from a still poorly Oseltamivir was never pivotal to Roche’s sure for orders from governments began to wane understood infectious disease. overall commercial success, which over the after the pandemic, the company sought to ramp past decade has been driven primarily by oncol- up interest among individual patients, funding Unlikely success ogy products developed with its US biotech an extensive public relations campaign as flu Since its launch in 1999, oseltamivir has gen- partner Genentech, which it fully acquired in Oseltamivir was never pivotal to erated cumulative sales in excess of $18bn 2009. In that year, which coincided with peak (£11bn; €13bn) for Roche. Half of the total oseltamivir sales, the drug was still only its Roche’s commercial success, which expenditure was by governments and com- fourth largest product, accounting for just 8% over the past decade has been driven panies around the world for stockpiles for of overall Roche revenues. primarily by oncology products

18 BMJ | 12 APRIL 2014 | VOLUME 348 ANTIVIRAL DRUGS

Roche’s sales figures show that since the pandemic sales of Tamiflu for seasonal use have increased

cases rose.3 It also targeted companies to buy Katrina in the US, a heatwave in France, and ity in patients admitted to hospital with H1N1 pandemic stockpiles. foot and mouth disease in the UK. influenza8 and an adviser to the UK government However, the pandemic was not a bonanza for Until 2009, when H1N1 emerged in Mexico, in the build-up to the pandemic, says: “I con- all drugs in the neuraminidase inhibitor class, neither public health officials nor Roche knew tinue to believe neuraminidase inhibitors are of which oseltamivir is the best known. Glaxo- the characteristics of the next flu pandemic. It a useful drug for patients with severe flu who SmithKline’s zanamivir (Relenza)—originally was impossible in advance to test oseltamivir’s are hospitalised. Cochrane only accepted ran- developed by Biota of Australia—was the first effectiveness on a still unknown virus, but a sci- domised control trials. If we had that sort of data such drug. Yet its cumbersome inhaled formula- entific consensus based on its use in seasonal we would give it primacy, but we don’t live in that tion was less appealing than the oseltamivir pill, flu suggested that the drug offered one of the world. We needed to use observational data.” and it has generated cumulative sales over the few interventions with the potential both to Since 2009, governments and public health past decade of less than $2.3bn. reduce the severity of infection and mortality bodies have paid little attention to revising BioCryst’s attempt to commercialise a third and to prevent disease. guidelines on recommended coverage levels for related antiviral drug—peramivir, which is It was seen as providing value in slowing future pandemics. Those that have been con- administered intravenously—have also been the growth of infection around the world and ducted tend to support coverage levels already slow. It has so far sold $23m of the product, flattening the surge in demand for treatment agreed in the late 2000s. while receiving $235m from the US government as a way to ease pressure on health systems Ironically, oseltamivir has proved chemically for development funding and licensing payments and buy time to develop vaccines. Even after extremely stable, with the result that many of from Shionogi of Japan and other pharmaceuti- the outbreak began, it took many months to the large remaining stockpiles around the cal partners totalling about $35m.4 understand that H1N1 was relatively benign. world have had their original five year shelf life In addition, testing oseltimivir in randomised extended by at least two years. Pressure to be prepared trials was considered by many at the time to New pandemic purchases on any significant Even these rivals’ sales—let alone most of those be ethically difficult. Given the high appar- scale are thus unlikely to take place before generated by Roche for oseltamivir—would ent demand at the time for oseltamivir, some 2016, when oseltamivir’s patents in most not have taken place without rising concerns campaigners argued that generic manufactur- countries expire and the drug can be offered over a flu pandemic. The drugs were originally ers should be allowed to produce it cheaply in more cheaply by generic manufacturers. Roche developed to treat seasonal flu, but there was larger quantities, with Roche paid a variable undoubtedly did extremely well from oseltami- little take-up other than in Japan. Health sys- royalty if the drugs were ultimately used to vir, far beyond its initial expectations. In the tems elsewhere were more sceptical about the treat patients.6 absence of robust clinical data that did not and drug’s value long before the recent scrutiny Looking back today, the strain of H1N1 iden- now cannot exist, taxpayers and public health of the Cochrane Collaboration questioning tified in 2009 proved far less fatal than plan- specialists may long continue to debate whether oseltamivir’s clinical value.5 ners had feared, and much of the stockpile the price paid was justified. What changed, and drove sales, was a ris- was never used. The Cochrane Collaboration Andrew Jack deputy analysis editor, Financial Times, ing focus on the risks of emerging infectious has thrown doubts on the value of oseltamivir, London, UK [email protected] diseases. At the start of the millennium, the based on a thorough review of the data from 5 Competing interests: I have read and understood the outbreaks of severe acute respiratory syndrome clinical studies of its use for seasonal flu. There BMJ Group policy on declaration of interests and have no (SARS) and then H5N1 “bird flu,” triggered con- remains heated debate about what the obser- relevant interests to declare. cern about the dangers of a new flu pandemic vational data, gathered during the pandemic This article is supported by a grant from the Open Society 7 Foundations. threatening humans. The scares coincided with years, tell us. Provenance and peer review: Commissioned; not externally growing pressure on policy makers to invest Jonathan Nguyen-Van-Tam, lead author of a peer reviewed. more in emergency planning, in the wake of Roche funded study that claimed oseltamivir References are in the version on bmj.com. a series of catastrophes including Hurricane significantly reduced mortality and morbid- Cite this as: BMJ 2014;348:g2524

BMJ | 12 APRIL 2014 | VOLUME 348 19 Improving General practice referrals in because they “introduce another general practice Tower Hamlets, east London, layer of administration, add had been creeping up for several costs, and deskill and undermine referral, Tower years, when in 2011 general GPs.” Instead the CCG developed Hamlets CCG, practitioners decided that action a package of interventions to was needed. But Victoria Tzortziou improve the management of ̻̻See the full shortlist at London Brown, a local GP and clinical common conditions in four http://thebmjawards.bmj.com commissioning group lead on specialties—musculoskeletal INNOVATION IN planned care and research, said disorders, dermatology, urology, that doctors wanted to avoid and ear, nose, and throat—and HEALTHCARE AWARD a referral management centre targeted referral behaviour. The

Bipolar education, MRC Centre for Neuropsychiatric It was the evidence that inspired Change Genetics and Genomics, Cardiff University doctors, nurses, and psychologists in Cardiff to develop the bipolar patient education programme. They knew that patients’ lives from the could be transformed by having a better knowledge of their condition. “There was a real desire among patients and front line professionals for an approach All the teams nominated that is not just focused on for this year’s award for Tackling chronic The surprising thing about Many people lacked a good innovation have been pain in people spending 90 minutes with understanding of their pain and someone with chronic pain— how to deal with it on benefits, Fit inspired by difficulties thoroughly examining them, For Work Team, talking to them about their intervention the DEAs said they encountered during their Leicestershire problems with debt and getting felt that they were working with everyday work, finds around, discussing possible completely different people— diagnoses and management that their clients were much Zosia Kmietowicz options—is the subsequent more engaged,” he said. reaction of their disability Hampton ran the “From Zosia Kmietowicz associate editor , BMJ, London, UK employment advisers (DEAs), pain to prospects” pilot, which [email protected] said Rob Hampton, a GP and looked at whether input from Competing interests: None declared. lead at the Leicestershire Fit a doctor, a pain management for Work service. “After this programme combined with

Safety incident reporting by families, Great In March 2013 Great Ormond Ormond Street Hospital, London Street Hospital began piloting a real time reporting system on its 14 bed renal ward. They wanted The Innovation in healthcare award is sponsored by to see whether involving families Greater Manchester Academic Heath Science Network and the BMJ Awards are sponsored by MDDUS. The of patients in reporting possible awards ceremony will take place on 8 May at the safety concerns might make the Park Plaza Hotel, Westminster. To find out more go to hospital safer. The novelty, in http://thebmjawards.bmj.com. today’s high tech environment, Provenance and peer review: Commissioned; not externally peer reviewed. was that the system involved 1 Smith DJ, Griffiths E, Poole R, di Florio A, Barnes E, Kelly MJ, et al. Beating bipolar: exploratory trial of a novel internet-based psychoeducational treatment Education for One of the challenges of treating desired therapeutic range for more for bipolar disorder. Bipolar Disord 2012;13:571-7. patients with atrial fibrillation than 70% of the time, was the 2 Poole R, Simpson SA, Smith DJ. Internet-based patients starting psychoeducation for bipolar disorder: a qualitative with warfarin is overcoming their inspiration behind the University analysis of feasibility, acceptability and impact. BMC warfarin for fear of harm because of the drug’s of Birmingham’s patient education Psychiatry 2012;12:139. 3 Clarkesmith DE, Pattison HM, Lip GY, Lane DA. atrial fibrillation, use as rat poison. That, and trying strategy. Educational intervention improves anticoagulation University of to maintain patients within the Together with patients, the control in atrial fibrillation patients: the TREAT team produced an intervention randomised trial. PLoS One 2013;8:e74037. Birmingham The intervention significantly that included a one-off group Cite this as: BMJ 2014;348:g2487 improved the proportion of session, an educational booklet, Headline Sponsor patients staying within the a self monitoring diary, and a therapeutic international worksheet. During the group normalised range during the first session patients were shown six months of treatment a DVD with clips discussing

20 BMJ | 12 APRIL 2014 | VOLUME 348 Auditing and discussing 2670 the following months led to a 15% to savings of £1m (€1.2m; referrals at practice meetings led to fall in referrals and a rise from 68% $1.7m) from reduced pathology a 15% fall in referrals to 79% in recording of suspected requests alone. The secret to diagnosis and history in referral the programme’s success? “GPs interventions were rolled out letters. Variation between practices were engaged from the start. across all 36 general practices in also levelled out, with the biggest Their enthusiasm to participate Tower Hamlets with a “referral reduction seen in dermatology was especially surprising,” said champion” in each of eight general referrals (interpractice variability Tzortziou Brown. “They saw [the practice networks. fell from 13.51% to 9.29%). The review of audits] as playing their Auditing and discussing 2670 programme has now expanded part rather than a punitive way of referrals at practice meetings over to other conditions, leading being judged.” giving medication,” said Ian skills as well as collaboration with have accessed online modules. There was a real desire among patients Jones, professor of psychiatry clinicians. Nearly all patients who completed and professionals for an approach that and deputy director of the MRC The dual delivery was essential, group tutorials (97%) said they is not just focused on giving medication Centre for Neuropsychiatric said Jones: “Not everyone likes a would recommend it to other Genetics and Genomics at Cardiff group course and eight group environment for learning, patients,1 2 and the package University. Instead they developed interactive web based modules and for them the online option has been embraced by the a holistic approach to increase called BeatingBipolar (www. of delivering the information has charity BipolarUK, the Welsh understanding of potential triggers beatingbipolar.org), both been a fantastic success.” government’s strategy for mental and the role of medication. providing information about To date 396 people across health, and staff training outlets There were two formats—a diagnosis and treatment and Wales have been through the in the UK, Northern Ireland, New 10 week psychoeducational encouraging self management group course and 3250 people Zealand, and Turkey. employment support, and a case five had never had individual manager can help people with physiotherapy and 23 had chronic pain get back to work. unmet health needs. Over four months in 2013 the Hampton said the findings programme assessed 31 people, highlight that GPs need more and so far five (16%) have found training in pain management a job, three are doing voluntary and that these patients need work (10%), and three (10%) are more time with a doctor than undertaking vocational training. the 10 minute consultation An unexpected finding, said allows. What was encouraging, Hampton, was that many people Hampton said, was the lacked a good understanding willingness of GPs to accept of their pain and how to deal and implement treatment with it. In addition, four out of recommendations.

Technology can sometimes be a prompting a conversation about communication. Being able to that families didn’t understand barrier to communication. Being their concerns. talk to someone face to face is what was happening to their able to talk to someone face to face Project manager Charlotte important,” she said. child, that poor communication is important Magness said this system was Of the families approached, between doctors and nurses chosen for its simplicity and 86% agreed to participate and 27 led to a delay in treatment, or a no computers or gadgets, just a ease of use for all families, families reported 33 problems, translator wasn’t available. Other laminated card with pictures and including those who could not with communication (31%) and wards at Great Ormond Street are colour coding. Family members speak English well and who have medication (21%) the largest now looking to adopt the scheme, could fill in the card at any time learning difficulties. “Technology categories of concerns. Problems as are safety advisers in the of day and take it to a ward sister, can sometimes be a barrier to with communication could mean Netherlands and Australia.

concerns about warfarin, the non-intervention group; advice from a cardiologist, P=0.035) and also significantly lifestyle recommendations, improved patients’ understanding and a mock consultation. A of the treatment as measured randomised controlled trial with a postal questionnaire after (the TREAT study) conducted the intervention. At 12 months from 2010 to 2012 showed that differences in the number in the the intervention significantly therapeutic range between the improved the proportion of patients groups were not significantly staying within the therapeutic different (76%v 70%; P=0.44),3 international normalised range suggesting that the intervention during the first six months of may need to be repeated to support treatment (76.2% v 71.3% in and reinforce behavioural change.

BMJ | 12 APRIL 2014 | VOLUME 348 21