How long does biomedical research take? Studying the time taken between biomedical and health research and its translation into products, policy, and practice

Background The time taken, or ‘time lags’, between biomedical/health research and its translation into health improvements is receiving growing attention. Reducing time lags should increase rates of return to such research. However, ways to measure time lags are under-developed, with little attention on where time lags arise within overall timelines. The process marker model has been proposed as a better way forward than the current focus on an increasingly complex series of translation ‘gaps’. Starting from that model, we aimed to develop better methods to measure and understand time lags and develop ways to identify policy options and produce recommendations for future studies. Methods Following reviews of the literature on time lags and of relevant policy documents, we developed a new approach to conduct case studies of time lags. We built on the process marker model, including developing a matrix with a series of overlapping tracks to allow us to present and measure elements within any overall time lag. We identified a reduced number of key markers or calibration points and tested our new approach in seven case studies of research leading to interventions in cardiovascular disease and mental health. Finally, we analysed the data to address our study’s key aims. Results The literature review illustrated the lack of agreement on starting points for measuring time lags. We mapped points from policy documents onto our matrix and thus highlighted key areas of concern, for example around delays before new therapies become widely available. Our seven completed case studies demonstrate we have made considerable progress in developing methods to measure and understand time lags. The matrix of overlapping tracks of activity in the research and implementation processes facilitated analysis of time lags along each track, and at the cross-over points where the next track started. We identified some factors that speed up translation through the actions of companies, researchers, funders, policymakers, and regulators. Recommendations for further work are built on progress made, limitations identified and revised terminology. Conclusions Our advances identify complexities, provide a firm basis for further methodological work along and between tracks, and begin to indicate potential ways of reducing lags. Electronic supplementary material The online version of this article (doi:10.1186/1478-4505-13-1) contains supplementary material, which is available to authorized users.

preventing 1 or more major forms of CVD and are less expensive. They should be preferred for first-step antihypertensive therapy.' (ALLHAT, 2002) 7. VALUE trial was international including UK but led from USA designed to 'test the hypothesis that for the same blood-pressure control valsartan [a Novartis ARB] would reduce cardiac morbidity and mortality more than amlodipine in hypertensive patients at high cardiovascular risk. ' (p.2022). Published in 2004 and funded by Novartis it found the main outcome of cardiac disease did not differ but reported: 'Blood pressure was reduced by both treatments, but the effects of the amlodipine-based regimen were more pronounced,' (Julius S et al., 2004) 8. ASCOT trial was an Anglo-Scandinavian trial led from Sweden comparing amlodipine with atenolol, a BB. Funded mainly by Pfizer, New York. Published in 2005. (Dahlof B et al., 2005. As stated in the paper, patient recruitment started in February 1998 and finished May 2000. Follow up was to be for 5 years. 'In Oct 2004, the DSMB recommended the trial be stopped on the grounds that compared with those allocated the amlodipine-based regimen those allocated the atenolol-based regimen had significantly higher mortality … between December 2004, andJune, 2005, the trial doctors recalled all patients for a final end-of-study visit.' The paper was published Sept 2005. 9. In the economic analysis for the 2006 guideline the cost data for the CCB used in the model was Amlodipine (p.70) because in all cases the prices of the most commonly used drug was included in the model. The health economics slightly favoured CCBs, but 'the more expensive brands are not likely to be cost effective for use in the NHS.' (p78) 10. The average time lag between publication of these 3 papers (2002,2004,2005) and guideline (2006) is much shorter than 'knowledge cycle time' average for guideline as a whole in Medical Research: What's it Worth (Buxton et al., 2008) of 9.5 years. But in that report it was also claimed there was an average 3 year gap between funding and publication, but in these 3 papers the gap between recruitment of first patient and publication is much longer (8 years, 7 years, 7 years respectively, notwithstanding the early end of the ASCOT trial and publication within four months of the end of data collection).
Then there are issues about the further update of the guideline in 2011 and more positive recommendations for CCBs because of the considerable reductions in price (which made it more cost effective for the NHS than doing nothing).
Apparently, all these studies have been important contributions to understand the mechanism of action of the calcium channel blockers. Late 1950searly 1960s First investigations into the pharmacology of calcium antagonists (unaware of their targets)

Discovery
Publications: Evans et al. (1958); Durbin and Jenkinson (1961) 1960s Fleckenstein and Godfraind demonstrated on arterial tissue the concept (and the mechanism of action) of calcium antagonists Discovery Publications: Fleckenstein et al. (1969); Godfraind and Kaba (1969) Late 1960searly 1970s The importance of Voltage-Gated Calcium Channels as target of the calcium blockers and their presence in heart cells is reported Discovery Publications: Reuter (1967Reuter ( , 1973 Late The paper of Faulkner et al. is not the publication of the studies but a publication citing the studies. In the original set of calibration points, we had a "cited on" calibration point. However, the last version of the minimum set of calibration points did not included this option.  11 October 1984 Dr Wells recommended the R&D group that the amlodipine besylate salt be substituted for the maleate salt in the commercial amlodipine tablet product This information is provided in the US Court Case FFCL223, par. 120.

End of 1984
Pfizer was conducting Phase II clinical trials of amlodipine maleate (using capsules and intravenous injections)

Human research: dosage/design (Phase II)
Possible citation on: Jackson et al. (1985) Possible citation on: Webster et al. (1987) This information is provided in the US Court Case FFCL223, par. 121.
I could not retrieve the Jackson et al. paper, so I am not sure if these Phase II clinical trials are cited in that paper.

1984-1986
Phase II double-blind placebocontrolled studies to investigate the efficacy of once daily amlodipine maleate in patients with mild to moderate hypertension.  13 13 October 1988 (1996) Web search using appropriate filters in: PubMed and Cochrane Library 14 1990 -A base-case analysis showed a slightly higher percentage of patients achieve hypertension control with amlodipine but at a substantially higher cost than with a generic diuretic (chlorthalidone)

Effectiveness/ postlaunch research
Publication: Ramsey et al. (1999) Web search using appropriate filters in: PubMed and Cochrane Library

Issues arisen in developing the timeline:
 [applicable to both amlodipine and olanzapine] If we keep both the "first nonresearch use in patients" and the "reimbursement/financial support" tracks this necessitates duplication of launch calibration point whenever we look at pre-NICE medicines  [applicable to both amlodipine and olanzapine] For pharmaceuticals, the "Intervention becomes standard of practice" could be questionable, as it is not clear how one can distinguish it from the "National policy announcement/ guidelines / advice" track  [applicable to both amlodipine and olanzapine] In some cases it was not possible to find the publication of a study/event and the citation in a publication about the study/event was used as calibration point instead. This can possibly introduce a bias in time lag estimation as the citation may occur several years after the event  [applicable to both amlodipine and olanzapine] Information about initial drug development (phase I and II clinical trials) is usually confidential. And, especially for "old" drugs, there are no online available public registry of trials  [applicable to both amlodipine and olanzapine] the US patents are relatively easy to retrieve, while there is no online available register for the UK patents  During a project workshop in January 2013 we agreed to consider also the approval date of the first generation calcium channel blocker (amlodipine is a third generation calcium channel blocker) to show the previous stages of research. Here I am considering the approval of the first generation calcium channel blocker in the Discovery track, however this might be incorrect and it might be the case where we should not consider it in the amlodipine timeline  When the potential publication of a study/event is not available, it is not possible to confirm that that publication actually refers to the study/event (e.g. first study in humans), so it cannot be stated for sure that the publication is a calibration point for the study/event Generic amlodipine was launched in the UK in March 2004. We do not have a track to capture this event, however it is an important event to take into account as it could have had a relevant impact on NICE cost-effectiveness and recommendation decisions.
This case study was conducted researching the online available information (both in published journals and in the grey literature) covering all the possible calibration points throughout the different tracks of the drug history, from early research to NICE clinical guidelines. This approach requires a deep knowledge of a drug research, development, launch and post-launch processes to identify the main events to be recorded; and involves both backward and forward tracing from the different calibration points. Valuable sources of information (which cannot be expected to be available in every drug case study) are the court cases, which reconstruct in detail the events related to the R&D of a drug.
There are advantages and disadvantages related to this approach. The main advantage is that this approach allows covering all the key stages of a drug history as it also considers the grey literature, which is a valuable (and sometimes the only) source of information for a drug commercial history (patents, regulatory application and approval, launch, national policy recommendation, uptake). When applied to more recent technologies, this approach potentially allows gathering more information, as it should be expected that drugs/interventions researched and developed in the Internet era have an extensive amount of records available online. One disadvantage of this approach is that it cannot be automated, as it requires seeking specific information manually. However, there are also economies of learning due to the recurrence of the main sources of information (e.g. the US Patent and Trademark Office or the FDA "Orange Book").
Some observations can also be made about the information availability and validity. In general, clinical studies at the earlier stages of a drug development (phase I and II) are not usually published in peer-reviewed journals or considered in research synthesis publications (an exception is represented by "more popular" drugs or drugs whose patent has been disputed in a court case). Moreover, how far peer-review publications record the start date of clinical studies varies, and practice is evolving. Furthermore ,the date of publication of the study may differ significantly from the date when the study took place. Furthermore, online datasets of clinical trials (e.g. clinicaltrials.gov) are relatively recent and therefore contain little information relevant to older drugs.
It should also be noticed that many different and staggered patents can be associated with a single drug. In this case, interesting information can be obtained investigating why different patents were granted. For instance, in the amlodipine case the new patent on the besylate salt testifies how the drug development can be complicated (and possibly delayed) because of formulation problems due to the stability of the salt (and not because the compound does not work).
It is also important to remember that NICE was set up in 1999. Therefore any measurement of lapsed time involving the publication date of a NICE guideline as calibration point should be considered as illustrative only if referred to interventions available before 1999.

Schizophrenia and olanzapine
Schizophrenia is a chronic, debilitating mental illness that appears during late adolescence or early adulthood and essentially lasts the lifetime of the patient. It is the most common major mental disorder in the UK, and directly affects one in 100 people at some point during their lives.
Scientists divide the symptoms of schizophrenia, which include delusions, abnormal thoughts and behaviour and lethargy, into 'positive' and 'negative'. The positive symptoms occur in the acute phase of the condition and include restless, noisy and irrational behaviour, sudden mood changes, inappropriate moods, hallucinations -often hearing voices -delusions and disordered thinking. The negative symptoms include tiredness, social withdrawal, physical slowness, lack of activity and interest in things and self-neglect, and usually occur in the chronic stage of schizophrenia. 3 Olanzapine is an atypical antipsychotic drug for the treatment of schizophrenia and bipolar disorder. It was originally sold by Eli Lilly and Company under the brand name Zyprexa® but today available generically. Olanzapine relieves the symptoms of schizophrenia (hearing, seeing, or sensing things that are not real, having mistaken beliefs, and feeling unusually suspicious) by correcting the imbalance of chemical substances which act on the nervous system in the brain.

Early Drug Treatment -Clozapine
Clozapine, the first atypical antipsychotic drug treating schizophrenia, was first marketed by Sandoz in Switzerland and Austria in 1972, in West Germany in 1974and in Finland in 1975 Prior to its launch, the only available drugs were typical antipsychotics, which only dealt with the positive symptoms of schizophrenia and had considerable side effects (acute extrapyramidal symptoms, EPS, were reported in 38.9% of patients). 5 Clozapine was considered a breakthrough innovation because it could treat both positive and negative symptoms of schizophrenia with no EPS.
However, it was soon found that a life-threatening blood disorder (agranulocytosis) was associated with clozapine treatment and the manufacturer voluntarily withdrawn the drug from the market in 1975. 6 This prompted pharmaceutical companies to start looking for a drug like clozapine, but without the same side effects.

Preliminary studies
Beginning in the autumn of 1974, Lilly UK started to research an atypical antipsychotic drug at its Erl Wood research facility. 7 The research team was led by research scientist Dr Jiban Chakrabarti, who first suspected that by altering the clozapine structure it would be possible to create a safer medicine. Research scientist Dr David Tupper and research chemist Dr Terry Hotten, who were working with Dr Chakrabarti, produced a novel class of compounds (thienobenzodiazepines), but only a few appeared to have an effect on psychoses. 8 In 1975, Lilly filed a patent application related to these compounds. The patent was issued by the US Patent and Trademark Office in 1978. 9 This patent led to a dispute in 1991 when the olanzapine US patent claim was filed.
Lilly spent the next four years developing one lead compound (flumezapine) through preclinical testing and initial safety testing in healthy human volunteers. 10 The first trial in actual schizophrenic patients started in the spring of 1982 but the first reports shown high toxicity level and the development of flumezapine was terminated.
After this failure, a team lead by Dr Hotten made and tested several alternatives, including the compound known as olanzapine, which was first synthesised in the UK on 29 April 1982. 11

Study in cell lines or animals (pre-clinical)
Lilly started testing olanzapine in beagles in 1983. The first study lasted three months, a one-year study followed later. 12 The tests indicated that olanzapine was sufficiently safe to be tested in human trials.

Phase 1 clinical trials
Phase 1 clinical trials in healthy human volunteers were conducted in 1986 and 1987 in Indianapolis, Indiana. 13

Phase 2 clinical trials
Clinical trials to test efficacy in actual patients took place in the UK and toward the end of 1989 demonstrated favourable results. 14 On 18 January 1990, the Research Management Staff agreed to "product commitment". 15

Patent application
On 25 April 1990, Lilly filed a patent application in the UK. The US patent application was filed on 23 April 1991. 16 The US patent application was rejected on 23 November 1991 declaring olanzapine unpatentable because it represented an obvious homologous substituent to the compound class patented in 1978 ("obviousness-type double patenting"). 17 Lilly responded to the rejection on 10 December 1992 18 arguing that olanzapine displayed an unexpected and significant superior toxicological benefit over the compound patented in 1978. 19 This claim was accepted on 17 December 1992 and the US Patent and Trademark Office issued olanzapine patent on 20 July 1993. 20 Phase 3 clinical trials In November 1993, Lilly obtained the results from the first pivotal clinical trial involving Zyprexa TM (LY170053, olanzapine), which demonstrated superior efficacy and safety to the existing therapies. 21 There have been four randomised double-blind clinical trials of olanzapine for treatment of schizophrenia prior to license. 22

Regulatory approval from FDA (US), EMA (EU), and/or MHRA (UK)
Although olanzapine was superior to the existing primary and secondary treatments for schizophrenia, Lilly acknowledged that several competitors had compounds in advanced stages of clinical trials. For this reason, Lilly developed a registration strategy which enable it to submit dossiers in 21 countries within several days of one another.
Nine months prior to the scheduled submission date (1 October 1995), the planning phase culminated with a global registration strategy. 23 The NDA and the European Marketing Authorization Application were submitted on the same day (22 September 1995) to United States FDA and the EMEA (now EMA) for the EU, respectively. Submissions were made to Australia, Canada, New Zealand, Norway, and Switzerland in the ensuing days. 24 EMEA approval was granted on 27 September 1996 and FDA approval on 30 September 1996. 25 It was launched first on October 1996 in 14 countries (including the UK) and then progressively in more countries over the next 21 months. 26 Uptake in the market The following graph shows uptake of olanzapine in primary care in England, measured in prescription items, following its launch in the UK in October 1996. Cited in:  1992-1997 Four double-blind pivotal studies, which compare olanzapine to placebo and/or haloperidol, are presented. The results suggest that olanzapine is as effective as haloperidol for positive symptoms and more effective than haloperidol for the treatment of the negative symptoms of schizophrenia.

Research Synthesis & clinical policy and practice to encourage uptake
Human research: efficacy (Phase III) Cited in: Beasley, Tollefson and Tran (1997).
Cited in:    NDA 20-592 (1996) Cited in:  18 October 1996 Begin of study F1D-MC-HGEH comparing olanzapine vs. placebo in the treatment of mania associated with bipolar I disorder This case study was conducted researching the online available information (both in published journals and in the grey literature) covering all the possible calibration points throughout the different tracks of the drug history, from early research to NICE clinical guidelines. This approach requires a deep knowledge of a drug research, development, launch and post-launch processes to identify the main events to be recorded; and involves both backward and forward tracing from the different calibration points. Valuable sources of information (which cannot be expected to be available in every drug case study) are the court cases, which reconstruct in detail the events related to the R&D of a drug.
There are advantages and disadvantages related to this approach. The main advantage is that this approach allows covering all the key stages of a drug history as it also considers the grey literature, which is a valuable (and sometimes the only) source of information for a drug commercial history (patents, regulatory application and approval, launch, national policy recommendation, uptake). When applied to more recent technologies, this approach potentially allows gathering more information, as it should be expected that drugs/interventions researched and developed in the Internet era have an extensive amount of records available online. One disadvantage of this approach is that it cannot be automated, as it requires seeking specific information manually. However, there are also economies of learning due to the recurrence of the main sources of information (e.g. the US Patent and Trademark Office or the FDA "Orange Book").
Some observations can also be made about the information availability and validity. In general, clinical studies at the earlier stages of a drug development (phase I and II) are not usually published in peer-reviewed journals or considered in research synthesis publications (an exception is represented by "more popular" drugs or drugs whose patent has been disputed in a court case). Moreover, how far peer-review publications record the start date of clinical studies varies, and practice is evolving. Furthermore, the date of publication of the study may differ significantly from the date when the study took place, and online datasets of clinical trials (e.g. clinicaltrials.gov) are relatively recent and therefore contain little information relevant to older drugs.
It should also be noticed that many different and staggered patents can be associated with a single drug. In this case, interesting information can be obtained investigating why different patents were granted. For instance, in the olanzapine case, the new patent testifies how previous research was not able to detect which compound within the patented class was the best candidate to be developed.
It is also important to remember that NICE was set up in 1999. Therefore any measurement of lapsed time involving the publication date of a NICE guideline as calibration point should be considered as illustrative only if referred to interventions available before 1999.

Background/definition
As reported in the main paper on this topic , ruptured Abdominal Aortic Aneurysms (AAA) caused about 6,800 deaths in England and Wales in 2000. Most were men. In men older than 65 AAA are responsible for 2.1% of deaths. An aneurysm is a localised widening, or dilation, of an artery. The blood vessel can burst, or rupture, because the vessel wall is weakened. The abdominal aorta is the largest artery in the abdominal cavity, and, as part of the aorta, it is a direct continuation of the descending aorta (of the thorax). Of the deaths from aneurysms in the abdominal aorta about half took place before the patient reached hospital, with the mortality rate for emergency surgery being between 30%-70%.
Ultrasound can reliably visualise the aorta, thus providing the possibility of detection through population screening at a size when rupture is unlikely to occur. The hope was that intervention at this stage could reduce the frequency of rupture, and so reduce mortality and the requirement for emergency hospital treatment.

Early applications of ultrasound
The account below demonstrates that there is no absolutely clear-cut starting point for a study of a screening programme for AAA. First, we briefly describe the development of ultrasonics in medicine (and that builds on the development of ultrasonics in other fields). Then we relate its early use in diagnosing AAA in the USA, and in a few cases in the UK. We also describe how at the same time several other methods were being developed for diagnosing AAAs and the potential of these for use in screening programmes was considered by several authors. Several key accounts in the USA and the UK identify a small study by Cabellon and colleagues in 1983 as being the start of population screening for AAA by ultrasonography (Cabellon et al., 1983).
The account of the use of ultrasonics in medicine started with its application in therapy rather than diagnosis (Woo, 2002). The first experimental uses in diagnosis seem to have been in the 1940s but at the First Congress of Ultrasound Medicine held in 1948 in Germany just two papers discussed it as a diagnostic tool, the rest were on its therapeutic use. Systematic investigations into using ultrasound as a diagnostic tool in the US took off in the late 1940s (Woo, 2002), with studies by Ludwig at the Naval Medical Research Institute in Bethesda (Ludwig and Struthers, 1949). In the late 40s and into 1950s Wild conducted a stream of research that became supported by the National Cancer Institute and, according to Woo (2000), pioneered ultrasound imaging as described in a paper in Science (Wild and Reid, 1952). In the UK important work on the use of ultrasound in diagnosis was conducted by Donald in the Department of Midwifery, University of Glasgow, and published in the Lancet (Donald et al., 1958). The acknowledgements in this paper credit the considerable contribution made by a local company that seconded an employee full-time to work with the university team and speeded progress.
Donald did report one example where he had used ultrasound to diagnose an AAA (Donald and Brown, 1961). This example does not feature in some of the main accounts of the history of using ultrasound for AAA diagnosis, but was reported by Bernstein et al. (1978).
Rather more frequently the study published in the USA in Jan 1966 by Segal et al. (1966) is described as the first case report of ultrasound for detection and size measurement of a patient with an AAA. In Oct 1966 Goldberg et al. published the findings from a small series of cases of ultrasonic measurement, and referred to Segal et al. as 'The only previous report of ultrasonic diagnosis'. Following that in the late 60s, and 70s and into the 80s there were a string of papers in the USA describing the use of ultrasound in the diagnosis of AAA when other factors indicated the likely presence of an AAA, and its use for measurement of AAAs to assist decisions about surgery. In an article published in 1977 Sutton and Garner claimed the method did not seem to be widely used in the UK, but they described how since Feb 1971 they 'assessed by ultrasound all patients referred to the X-ray department, St Mary's Hospital, with a diagnosis of aortic aneurysm. ' (1977, p.741). In Glasgow McGregor and colleagues (1975) reported using ultrasonography in the diagnosis of AAA: '59 patients with prominent abdominal pulsation were examined and aneurysm was demonstrated in 22.' (p.133).
Meanwhile various other approaches, beyond physical examination, were being explored to assist diagnosis and treatment of AAA. For six and a half years between January 1964 and May 1970, Schilling and colleagues studied 2663 subjects; 1517 of the total were examined by lateral abdominal roentgenography annually on 2 -7 occasions. They concluded: 'inasmuch as many aneurysms are asymptomatic, routine radiologic surveys are indicated at age 55 as operative risk is less enhanced by degenerative disease.' (Schilling et al., 1974). A range of papers describe various techniques to diagnose AAAs, including descriptions of radiography, angiography and Computed Tomography -CT. In a chapter published in 1982, and based on a conference on AAAs, one of the major US teams stated: 'Improving ultrasonic and CT scanning technology should permit the risk-free and accurate diagnosis of abdominal aneurysms on a routine basis in an outpatient setting.' (Bernstein et al., 1982, p. 230). A team at the Walter Reed Army Medical Center in Washington led by Dr Norman Rich had been attempting to develop ways to better evaluate AAAs, but had abandoned a prospective study set up in the later 1970s because the angiography was sometimes misleading (Rich et al., 1982).
A member of that team, Dr Silverio Cabellon, moved to the William Beaumont Army Medical Center in Texas and decided CT scanning would be too expensive, but that he could use the new ultrasound equipment available to him in a prospective way to routinely diagnose the presence or absence of AAAs in all patients with peripheral atherosclerotic vascular disease who met certain criteria. He reported on all 73 patients examined in the first 3 months (Cabellon et al, 1983). This study has been identified in the US and the UK as the starting point for AAA screening by ultrasound. (see Scott et al. in MASS application 1996, Wolf et al,1995Lederle, 2003;Lederle, 2008) Key issues in the timeline of the development of AAA screening and analysis of possible time lags The timeline from 1984 onwards is dominated by the work of one team in the UK (the Chichester team led by Alan Scott and their collaboration with the Health Economics Research Group (HERG), Brunel University, and the MRC Biostatistics Unit at the University of Cambridge.) The two RCTs conducted by this team constitute over 75% of the weight of evidence in the international reviews on this topic. Key issues from this time line include the rapid start of the first RCT (the Chichester study) in 1989 even before the pilot Community cohort study being conducted by the team had been completed: the pilot had crucially demonstrated the feasibility of the methods. Likewise, an application for the major MASS trial was submitted before the Chichester study had been completed, and the proposal for the MASS study made clear that the trial was intended to inform the policy on screening. There was a time lag here and a revised proposal had to be submitted to address issues including the organisation of the proposed study to ensure it was large enough to be a powerful study, and the desire to wait until the results were known from a trial about the benefits of surgery on small aneurysms. (This illustrates a major factor with screening which is the desirability of fully understanding the appropriate way to treat those identified as having the condition for which the screening is being conducted. This is especially important in a field where the treatment option itself carries a risk).
There was again an attempt to reduce any time lags because the UK National Screening Committee (NSC), which had previously made clear in documents that it was waiting for the results from the MASS trial (UKNSC, 2001), requested a presentation from the MASS team before the publication of the findings. There was considerable discussion about how best to implement any screening programme, but the NSC's decision in principle to recommend going ahead with a programme was made in 2005 (UKNSC, 2005) even before a Cochrane systematic review was published in 2007 (Cosford et al., 2007). This decision is being considered as the equivalent of regulatory approval.
There was further analysis, planning and discussions about aspects of implementation before the Ministerial decision to go ahead with the policy was announced in early 2008 (Brown, 2008). The subsequent analysis necessary for the Impact Assessment (DH, 2008) that accompanies government spending decisions drew on the Cochrane review as well as the MASS study. The impact assessment estimated the health gain that would arise from full implementation and suggested it could prevent 2,800 deaths a year. It also estimated the net benefit of the programme, based on the quality adjusted life years to patients gained, at £3,884.1m over 20 years. The programme was implemented in England in a phased way between 2009 -2013. The continuing findings from the follow-up studies informed aspects of the implementation of the screening programme. Implementation of the NHS AAA screening programme in England was successfully completed in Spring 2013. According to the programme's web site: 'The NHS AAA Screening Programme covers the whole of England. The programme is delivered by 41 local screening services which together offer screening to around 300,000 men every year during the year they turn 65….AAA screening programmes will be fully implemented across Northern Ireland, Scotland and Wales by the end of 2013.' (NHS Abdominal Aortic Aneurysm Screening Programme -Frequently Asked Questions) There are also indications that the insistence by the MRC that the study be revised to make it more powerful might eventually have helped the translation process because the quality of the study was referred to in various documents as being one of the reasons why the research was translated into policy. Furthermore, an important contextual point that helps explain the international significance of the UK study was set out in the proposal to the MRC: 'screening studies with a control population are not possible in countries with a predominantly private health care system.' (In the USA where there were no RCTs, there was an even more rapid introduction of policy with the US Preventive Services Task Force systematic review in 2005 drawing heavily on the MASS study (USPSTF, 2005;Fleming et al., 2005). Following that The Screening for Abdominal Aortic Aneurysms Very Efficiently (SAAAVE) Act was passed in 2006 and allowed Medicare to offer screening from January 2007 (Lederle, 2008). There is not, however, universal provision of AAA screening for those reaching 65).
Finally, it is possible to argue that the researchers, the MRC and the NSC combined in this case to play the role that a company plays in the development and uptake of a drug -proactively driving the research and its translation through to ultimate uptake and considerable health gain.

Study(ies)/Events Track Calibration point Comment
July

1983-Nov 1983
First reported (uncontrolled) study of ultrasonic screening for AAA involving consecutive patients: Cabellon et al.

Human research
Start of patient recruitment: July 1983; End of patient recruitment: Sept 1983; Publication: Nov 1983: (Cabellon et al, 1983) Generally described as the first ultrasonographic screening paper: Scott et al application to MRC (referenced here as MRC, Jan 1996); Wolf et al, 1995;review by Lederle, 2003 and. Study was small involving just 73 patients, and selected population, but it was a sequence of all patients diagnosed at the Medical Centre as having atherosclerosis which was viewed as a major risk factor for AAA.

1983-6
Uncontrolled study of two groups of 100 hypertensive patients  Kim et al, 2007; (7 year data) Thompson et al, 2009; (10 year data) final follow-up: Thompson et al, 2012;(13 year data) Proposal turned down (as described above) partly because results of surgery trial were not yet available; Jan 1996 meeting of MRC's Health Services and Public Health Research Board asked for the proposal to be revised and resubmitted.
UK National Screening Committee (NSC) asked for an early presentation of findings, prior to publication.
Contributed 65.9% of the weight in the US Preventive Services Task Force review (Fleming et al, 2005) (Fleming et al, 2005) Early comment by Spencer et al (2000) that selective screening for AAA using easily recognisable risk factors is feasible, but not worthwhile as approx. 25% of the clinically significant cases would be missed.
Research review Publication: Lederle, 2003. The 4 studies described above were the only ones included: Chichester; MASS; Western Australia; Denmark. MASS the largest. Also Cabellon et al (1983) is the first ultrasonography study mentioned.

2005
First internationally recognised review conducted for a government agency: US Preventive Services Task Force (Fleming et al, 2005) Research Aneurysm highlights the quality of the cost effectiveness analysis in MASS. It states: 'we believe that the detailed micro-costing approach used in the MASS CEA, as well as its use of probabilistic sensitivity analysis, mitigated its being set outside the United States (it was conducted in the United Kingdom) and justified a "good" quality rating.' (USPSTF, 2005(USPSTF, ) 2005 UK National Screening Committee recommendation to ministers for introduction of a population screening programme in principle.

Regulatory approval
Regulatory approval: NSC, 2005: 'Screening of men aged 65, with the offer of a single test being made at that age, can be recommended in principle as a programme that meets the criteria and standards of the National Screening Committee. The configuration of services is a critical issue to be considered further before any implementation of new screening programmes.' (UKNSC, Nov, 2005) The decision in principle is here viewed as the equivalent to regulatory approval, but the decision to introduce a national population screening policy has to be taken in the UK by ministers, and further work was required. There is evidence from the files of considerable liaison (dissemination) between the research team and the NSC leading up to and after the decision by NSC.  (2009) Guideline identified 4 RCTs, ie the 4 described above. Recommendation: 'One-time ultrasound screening for AAA is recommended for all men at or older than 65 years. Screening men as early as 55 years is appropriate for those with a family history of AAA.' The Level of recommendation was described as 'Strong', the Quality of evidence described as 'High'.
As early as Jan 2004 there had been advice in form of Consensus Statement from 3 American vascular societies recommending AAA screening (Kent, 2004)  In this case it was possible to make considerable progress in identifying the detailed steps of the timeline, and assess the time elapsed as being 26 years from the first screening study to the intervention becoming standard practice. A full range of case study techniques was successfully deployed to make this progress: documentary and archival analysis (which was greatly aided by a leading member of the MASS team, Prof Martin Buxton, being a member of the time lags project team); a Google search; a review of relevant publications; interview with a key player, Dr Silverio Cabellon, who conducted the first AAA ultrasound screening study.
For AAA it was possible to identify start points/start of patient recruitment for some key projects and these were very important for understanding the nature of the time line, eg the case study showed how the first RCT started before the pilot study had published its findings, so it would have been misleading had the case study relied solely on publications. Also for the MASS study, the information about the date the proposal was originally submitted and then re-submitted was very useful and would have been hidden by a reliance solely on the publication dates of the first RCT, which was described as a pilot for the bigger RCT.
AAA was one of the cases where it was also possible to gather information about some important markers listed in Trochim but excluded from our final collective list. In some instances these proved to be very useful for understanding time lag issues, for example in MASS the details about the end of patient recruitment enable us to see that the study involved a relatively short period of patient recruitment, but a longer follow-up period.
We might need to consider if the scope or length of some of the studies was sufficient, for example while the pharmaceutical case studies described earlier identified important time lags by going back into the Discovery track, this part was missing from the AAA study which only focussed on a shorter span and so did not analyse possible delays that were touched on in the background to the case study.
The matrix allows the presentation of the data in a way that clearly demonstrates what may be a feature of the time lines for screening research, ie a relatively long research period, which is necessary to conduct thorough studies of the potential benefits of introducing screening policies, followed by a reasonably rapid translation into policy that is then implemented. In this case, however, the usual phases of human research that are relevant in a pharmaceutical intervention do not translate neatly into the types of studies conducted for a screening intervention.
There were a few aspects, such as the first small uncontrolled study in the UK by Twomey et al (1984 and1986), where some details are missing from this account and it might have been useful to undertake further interviews. However, this case study concentrated on a specific intervention and was able to work along the time dimension reasonably comprehensively by focussing on the details of the events and studies that were key to understanding the time elapsed, and the causes of the small delays that did occur.

Time lags between conducting medical research and its translation. Case Study 4: Smoking reduction NARRATIVE ACCOUNT OF BACKGROUND/ DEFINITIONS AND KEY ASPECTS OF THE INTERVENTION'S DEVELOPMENT
While this case study focuses on smoking as a cause of cardiovascular disease, much of the evidence and policy reviewed relates to the relationship between smoking and lung cancer. This is because it was this relationship with cancer which led to the majority of policy responses and interventions, and as such the evidence relating to cancer is a more appropriate point to start from when exploring the time lag. For comparison, however, the timeline does note key research, reviews and statements relating to cardiovascular disease.
Although writers had speculated about a link between tobacco and cancer since the 1850s, it was not until 1912 that Isaac Adler first proposed that smoking (as opposed to tobacco dust) might be a cause of lung cancer. The first links between tobacco and arteriosclerosis were also suggested in the late 19 th century. The number of deaths from lung cancer increased dramatically in the first half of the 20 th century, leading to the MRC organising a conference in 1947 to examine possible causes. The suggestion at this conference that tobacco, and in particular cigarette smoking, might be a factor led to the MRC funding a case control study to look at the link between smoking and lung cancer (Doll & Hill, 1950). This study, along with four US studies published the same year, all concluded that cigarette smoking was likely a causal factor in the development of lung cancer.
Building on these studies, and realising that a different kind of evidence was needed to make a conclusive case, Doll and Hill began a prospective cohort study of 40,000 British doctors, tracking their smoking habits and investigating whether it was possible to predict lung cancer risk. Preliminary results in 1954 suggested an association between smoking and lung cancer, as well as some effect on coronary thrombosis (Doll & Hill, 1954), while in the same year early findings from a large US study indicated a higher death rate among smokers from both lung cancer and coronary artery disease (Hammond & Horn, 1954). Although it had first been suggested back in 1934 that smoking might contribute to coronary thrombosis, and further evidence of a link with cardiovascular disease had emerged during the 1940s, this had not been conclusive.
Further findings from the British Doctors Study were published in 1956, and even at this early stage the authors were able to conclude a death rate from cancer that was 20 times higher in smokers compared with non-smokers. At the same time, evidence was also beginning to build on the nature of tobacco, with the discovery of carcinogens in tobacco smoke and evidence on the effects of tar in causing tumour growth in animals. In 1957, both the MRC and the US Surgeon General declared a causal link between cigarette smoking and cancer, and in the following years, influential reports from the US Surgeon General and the Royal College of Physicians summarised the evidence and recommended a variety of restrictions on sales and marketing.
Throughout the 1960s and 1970s, a range of advertising restrictions were introduced (with a UK ban on TV advertising of cigarettes enforced in 1965) and various anti-smoking campaigns implemented, focused both on lobbying for smoke-free environments (for which public support was growing) and on encouraging people to stop smoking. During this time the evidence on smoking harms had also continued to build, both from the British Doctors Study's ongoing analysis and more widely, with an increasing emphasis on the potential effects of environmental tobacco smoke. The first epidemiological evidence linking passive smoking to lung cancer was published in 1981, and in 1986 a number of reports concluded a causal link, leading to proposals for further restrictions on smoking in public places (e.g. Froggatt, 1988). Similarly conclusive epidemiological evidence on the role of passive smoking in cardiovascular disease was published in 1991 (Glantz & Parmley, 1991 At the same time as evidence was mounting on the effects of passive smoking, there was also an increase in the availability of smoking cessation services to individuals (initially through organisations such as ASH and QUIT, then on the NHS from 1998) and moves to further restrict tobacco advertising. In 1990 the European Parliament voted in favour of a total ban on advertising, which led to a European Commission proposal for a ban in 1991, and the eventual adoption of an EU Directive banning advertising and sponsorship in Member States in 1998. This Directive was overturned by the European Court of Justice three years later as its implications were deemed beyond the EU's powers, but a more restricted Directive replaced it in 2002. In the UK, the 1992 Smee report summarised the evidence on the effects of tobacco advertising on consumption and the 1994 action plan to reduce smoking included measures to restrict advertising. In 1996 Guernsey became the first government in the British Isles to agree a complete ban, before a similar bill was passed in the UK parliament in 2002. Most recently, the UK Government launched a public consultation on plain packaging of cigarettes (2012) (Huchard, 1893) 1898 Hermann Rottmann, a medical student, in Würzburg proposed that tobacco dust-not smoke-might be causing the elevated incidence of lung tumours among German tobacco workers Discovery Source: (R. N. Proctor, 2012) 1904 Erb (1904) finds that 25 out of 45 patients with intermittent claudication were heavy smokers Discovery Publication: (Erb, 1904) 1908 Buerger (1908) describes a rare form of vascular disease in young people and notes that it rarely occurs in non-smokers. These findings are confirmed by several others over the following 30 years, but the occasional diagnosis of the disease in non-smokers was considered to rule out Discovery Publication: (Buerger, 1908) Source: (Doll, 1998) 54 smoking as a cause.
1912 Isaac Adler first links smoking with lung cancer (rather than tobacco dust), but states that this is only a possible cause and notes that there is not sufficient evidence (Adler, 1912) Discovery Publication: (Adler, 1912) 1934 First suggestion that smoking might be responsible for an increase in coronary thrombosis Discovery Publication: (Howard, 1934) 1939 Müller (1939) publishes first case control study, comparing retrospectively smoking behaviour of patients who had died of lung cancer with healthy controls. He concluded that "the extraordinary rise in tobacco use" was "the single most important cause of the rising incidence of lung cancer".
1940 English et al. (1940) compare (i) the habits of 1000 patients with coronary thrombosis to 1000 matched controls, and (ii) the incidence of coronary disease in 1000 smokers compared with 1000 non-smokers. They conclude that smoking tobacco probably has "a more profound effect on younger individuals owing to the existence of relatively normal cardiovascular systems, influencing perhaps the earlier development of coronary disease".

Discovery
Publication: (Schairer & Schöniger, 1944) Details in Doll (1998) (Wassink, 1948) 1950 Five key papers published, all describing case control studies showing a link between smoking and lung cancer: Doll and Hill (1950)  Human research

Start of data collection (Oct 1951):
Detailed in Doll and Peto (1976) 1953 Ernst L Wynder finds that painting cigarette tar on the backs of mice creates tumours. This is the first biological link between smoking and cancer.

Discovery
Publication: (E. L. Wynder, Graham, & Croninger, 1953) 1954 Preliminary results of the British Doctors Study published. Doll and Hill suggest an association between smoking and lung cancer and a "significant adjuvant effect" of tobacco on coronary thrombosis.

Discovery
Publication: (Doll & Hill, 1954) 1954 Preliminary results of a large US study of over 187,000 men indicate a higher death rate overall, from coronary artery diseases and from lung cancer among smokers than among non-smokers. The authors also concluded that "probably nicotine is at least partially responsible for the findings in relation to diseases of the coronary arteries." Discovery Publication: (Hammond & Horn, 1954) on the association between smoking and lung cancer.
1955 Cooper and Lindsey (1955) demonstrate existence of carcinogens in cigarette smoke Discovery Publication: (Cooper & Lindsey, 1955) Mid-1950s In the US, individuals begin to sue tobacco companies for damages 1956 Doll and Hill publish first results of the British Doctors Study, demonstrating a death rate from lung cancer 20 times higher in those that smoke than those that don't (Doll & Hill, 1956) Human research Publication: (Doll & Hill, 1956)  Publication: (Hirayama, 1981) cohort study in Japan. Concluded lung cancer incidence higher among non-smoking women married to smokers than those married to nonsmokers. 1983 Surgeon General's report states that "smokingrelated cardiovascular disease is estimated to account for more deaths than any other smokingrelated disease, including cancer", noting that smokers' death rates from coronary heart disease were 70% higher than for non-smokers.

Research review & synthesis
Key review: (Froggatt, 1988(Froggatt, ) 1988 In the US, a court awards damages against a tobacco company for a lung cancer death. Publication: (Doll, et al., 2004) expectancy by around ten, nine, six and three years, respectively. (Doll, Peto, Boreham, & Sutherland, 2004) 2004 Government adviser Derek Wanless publishes 'Securing Good Health for the Whole Population', which recommends banning smoking in work places among other things. This is echoed by the CMO in his 2003 annual report, which concludes that this would bring a net benefit to society of £2.3-2.7bn annually.

Research review & synthesis
Key review: (Wanless, 2004) 2004 A government white paper follows proposing a ban in the majority of workplaces and public places (but short of a total ban).  Interventions that are commercially developed are not included (e.g. drugs, stop smoking aids).  Evidence on harms other than cardiovascular disease (the focus of the study) and cancer (the main driver of policy and interventions relating to smoking) was not included to keep the scope manageable.
 While the case study includes research evidence internationally, it focuses on intervention in the UK. This is mostly restricted to national level policies, campaigns and activities, as it was not possible to capture everything that has taken place locally or regionally.
 Discovery/human research is defined as follows: harms evidence is classified as discovery unless a comparison with a non-smoking group is used -the logic being that in these cases not smoking is effectively the 'intervention'. Any research looking at a specific intervention is classified as human research.
 Subcategories within human research have not been used, as these were not easily definable in this case study.
 Evidence on effectiveness of the various 'interventions' is rarely included, as in an attempt to constrain the scope, the case study focus has been on evidence of harms and details on implementation. In some instances, the evidence on effectiveness is either not specific to smoking/public health (e.g. on effects of taxation) or only emerged following the implementation of an 'intervention' as its effectiveness was assessed. In this latter case, effectiveness evidence will have had little/no effect on the time lag.
 An important aspect of this case study is awareness raising -e.g. campaigns, media coverage, information provision. This does not fit easily into our 'track' structure. Campaigns can be classed as interventions, but this is less clear for events such as a TV documentary on the tobacco industry or a statement by a lobbying group.
 In some cases, it is difficult to distinguish between tracks e.g. in cases where the national policy is the intervention, such as increasing tobacco duty. For this reason some 'events' have not been classified into tracks.
 Calibration points -the following definitions are used: o Announcement about national policy -a government statement of intent (e.g. commit to reducing smoking, or to implement a ban in future) o Guidelines issued -when non-compulsory guidelines or codes of practice are launched (e.g. around advertising) o Intervention becomes standard practice -when a policy or regulation comes into force and is compulsory (this is the same as a national policy being implemented, but by definition it becomes standard practice immediately if it is a law)

Definition of CBT
It is generally accepted that 'CBT' refers to a 'family of allied therapies that draw on a common base of behavioural and cognitive models of psychological disorders and utilise a set of overlapping techniques' (Churchill et al., 2012). The full NICE guidelines on depression define CBT as: … time-limited, structured psychological interventions, derived from the cognitive behavioural model of affective disorders and where the patient:

 works collaboratively with the therapist to identify the types and effects of thoughts, beliefs and interpretations on current symptoms, feelings states and/or problem areas  develops skills to identify, monitor and then counteract problematic thoughts, beliefs and interpretations related to the target symptoms/problems
 learns a repertoire of coping skills appropriate to the target thoughts, beliefs and/or problem areas.
In most individual trials of CBT, the manual used was Beck's Cognitive Therapy of Depression (1979) which advocates 16 to 20 sessions for treatment and relapse prevention work.

The development of CBT to treat depression
Throughout the 1960s and 1970s A.T. Beck wrote extensively on the connection between depression, cognitive distortions and the use of CBT to treat depression, which led to the first RCT being carried out in 1977 (categorised as a phase III study in the timeline), and CBT being developed for the treatment of depression in 1979. Between 1977 and 1998 78 RCTs were conducted in this area, and in 1989 the first meta-analysis was published.
However, it wasn't until 2000 that a series of cost-benefit studies were undertaken, which is thought to have unlocked funding to roll out the implementation of CBT across the NHS. The full NICE depression guidelines (2009), describe the development of CBT as follows: the development work took place in the 1960s and 1970s; the manual was published in 1979 (Beck et al., 1979); the first RCTs were published in the late 1970s and early 1980s (Rush et al., 1977;Kovacs et al., 1981;Rush et al., 1981); the first meta-analysis was conducted in 1990 (Robinson et al.,1990); and the effectiveness and cost-effectiveness studies have only started to emerge in the last decade Byford et al., 2003;Scott et al., 2003). In summary, over the past 50 years there has been a significant expansion of theories and therapies for depression. However, only a relatively small number of these therapies have travelled the full empirical road and demonstrated that they are efficacious and can be cost-effective treatment options for the NHS.
The use of CBT to treat depression was recommended in the Department of Health's 'Treatment Choice in Psychological Therapies and Counselling' in 2001 and in 2004 the first NICE guidance on depression was published which also recommended CBT for the treatment of depression. However, despite these recommendations CBT was not widely available given the lack of qualified CBT therapists. In 2006 the Improving Access to Psychological Therapies (IAPT) initiative was launched which from the outset 'recognised that a national shortage of CBT practitioners, who are skilled in helping people recover from depression and anxiety disorders, was the core deficiency preventing routine NHS delivery of the NICE guidelines' (Improving Access to Psychological Therapies, 2006). As part of the IAPT initiative in 2008 Health Secretary Alan Johnson announced an extra 3,600 therapists would be trained to provide CBT treatment on the NHS at a cost of £173 million per year from 2010. The scope of CBT and computerised CBT (CCBT) was broadened in the NICE 2009 depression guidelines which suggest that CBT (either alone or in combination with medication) should be used for mild, moderate and severe cases of depression. The 2004 guidelines had stated that there was insufficient evidence to introduce CCBT technology into the NHS.
In 2011 the Department of Health released 'Talking Therapies: A four-year plan of action' which underscored how important the economic case was in rolling out the use of CBT across the NHS. It stated that 'the evidence that proved CBT is as effective as medication in helping people with depression and anxiety disorders -and better at preventing relapse -led to the economic case that secured annual funding to begin the national roll-out in the three years to March 2011. Key to the economic case was an argument that effective therapeutic interventions combined with employment support could reduce the numbers of people on sick pay and benefits'.

TIMELINE
All dates in the period/date column refer to publication dates. Publication dates have also been used to draw the matrix.   Reason and Emotion in Psychotherapy, Lyle Stewart: New York

Reviews:
Cited in Cochrane Review Protocol (Churchill et al., 2012) Guidelines: Cited in full NICE depression guidance (2009) 1977-1998 78 randomised controlled clinical trials regarding the use of CBT and depression published (see Gloaguen, et al., 1998) Narrative

N/A 1960s and 1970s
Beck wrote extensively on the connection between depression, cognitive distortions and the use of cognitive therapy as treatment  New York: Springer, 1977 First RCT of CT in depression carried out (Rush et al., 1977) n=41 (n=19 cognitive therapy, Human researchefficacy (phase III)

Publications:
Rush, A.J., Beck, A.T., Hollon, S.D. et al., (1977) Comparative efficacy of cognitive therapy and pharmacotherapy in the treatment of depressed This case study was developed through consulting a range of documents identified through literature searches in PubMed and Google Scholar, and analysing the references of each document. Therefore, the sole method employed in developing the case study was mining literature. This process was made more difficult given that a Cochrane review on the use of CBT and depression has not been published (although there are Cochrane reviews on the periphery of this subject). This may be something to bear in mind regarding recommendations to automate the process. It should be noted that this case study does not encompass all RCTs or studies more generally undertaken on the subject of CBT and depression (as there are a very large number). Rather, it attempts to identify key studies throughout the narrative. All RCTs undertaken after CBT was recommended to treat depression in NICE guidelines in 2004 have been categorised as 'effectiveness/post-launch research' along with cost-benefit studies which took place prior to 2004. Although RCTs are sure to continue beyond 2008, this stream stops based on the most recent relevant RCT in the Cochrane review protocol, which was the Stiles study published in 2008.
A key obstacle in developing the case study was identifying a range of calibration points such as proposal submitted, first patient recruited and dates of data collection. In most cases this information was not given in papers, and therefore where it was given it could not be used for reasons of consistency. If other case studies have used calibration points other than publication dates, the time lags between this work and others may not be comparable.
More generally, the matrix worked well in providing a visualisation of CBT's development. However, it did not provide any insight regarding the cause of time lags in CBT's development -particularly given that it does not capture negative events. If the matrix were to capture and clarify causes it would need to capture general narrative and context as well as just events -e.g. to show the lack of trained CBT practitioners in preventing the widespread adoption of CBT.

Definition of CBT
It is generally accepted that 'CBT' refers to a 'family of allied therapies that draw on a common base of behavioural and cognitive models of psychological disorders and utilise a set of overlapping techniques' (Churchill et al., 2012). The full NICE guidelines on depression define CBT as: … time-limited, structured psychological interventions, derived from the cognitive behavioural model of affective disorders and where the patient:

 works collaboratively with the therapist to identify the types and effects of thoughts, beliefs and interpretations on current symptoms, feelings states and/or problem areas  develops skills to identify, monitor and then counteract problematic thoughts, beliefs and interpretations related to the target symptoms/problems
 learns a repertoire of coping skills appropriate to the target thoughts, beliefs and/or problem areas.
In most individual trials of CBT, the manual used was Beck's Cognitive Therapy of Depression (1979) which advocates 16 to 20 sessions for treatment and relapse prevention work.

The development of CBT to treat schizophrenia
A.T. Beck first applied CBT techniques to schizophrenia in 1952, although it had little impact as schizophrenia was thought to be incurable. Beck himself said that after publishing this paper he forgot about it and went on to focus on depression instead. This was reinforced by Freud's announcement in 1957 that psychoanalysis cannot be used as a treatment for schizophrenia as people with schizophrenia cannot form proper attachment to a therapist. From 1979 until the late 1980s very small studies were carried out (some only involving 1 patient), assessing how CBT could be applied to schizophrenic patients. This caused A.S. Bellack, president of the Association for the Advancement of Behavioural Therapy to describe schizophrenia as 'behaviour therapy's forgotten child' in 1986. (It is worthy of note that during this time studies for the use of CBT and depression were much larger and I have categorised them as phase III studies as opposed to phase II studies). However, throughout the 1980s there had been consistent findings that family environments were strongly influential in determining the recurrence of schizophrenic symptoms and a group of psychologists in the UK using rehabilitation approaches had also started taking cognitive approaches in understanding rehabilitation.
In 1988 David Kingdon was working on the Nottingham Neurosis Study (Tyrer et al, 1988(Tyrer et al, , 1990 in which cognitive therapy was employed by a group of patients with neurotic disorders. Kingdon realised that the techniques he was using with patients who had psychoses could be described as a variant of cognitive therapy. This prompted him and colleagues to define the techniques used, incorporate additional components, and refine them with further study of the literature. Throughout the early 1990s there were several developments of CBT interventions for psychotic patients and by 1996 there was enough evidence to start generating RCTs. From 1996-2002 there were 6 key RCTs with promising results. However, the first Cochrane review on the use of CBT to treat schizophrenia was published in 2002 and found that 'trial-based data supporting the wide use of cognitive behavioural therapy for people with schizophrenia or other psychotic illnesses are far from conclusive' (Cormac et al., 2002 p.2).
In 2002 the first NICE guidance was published which stated that CBT (and family interventions) should be used in the treatment of schizophrenia. This was followed in 2004 by the American Psychiatric Association recommending 'cognitive behaviourally oriented psychotherapy' to treat patients with schizophrenia, and in 2005 by the Canadian Psychiatric Association recommending CBT for treatment-resistant schizophrenic patients. The reason for accelerated approval in the UK may be because in the US the separation of psychology and psychiatry has prevented a multidisciplinary approach to treatment -which facilitated the development of CBT for the treatment of schizophrenia in the UK. Moreover, research funding in the US is more widely available in biological psychiatry which has contributed to the prominence of biological psychiatrists in the US.
In 2006 the National Forum for Assertive Outreach Annual Conference surveyed attendees to find out what the barriers were to the implementation of psychosocial interventions in their area.
Respondents reported a lack of organisational investment, the structured nature of CBT, caseload issues, medication issues, application to people with sensory impairment, staff apathy and staff burnout as some of the barriers to the implementation of CBT for psychosis. In response to the lack of organisational investment, Improving Access to Psychological Therapies was launched by the Department for Health in the UK.  . These findings may have come too late to have a real impact on the uptake of CBT, given the money that has already been invested into rolling out the treatment across the NHS.

TIMELINE
All dates in the period/date column refer to publication dates. Publication dates have also been used to draw the matrix.  (Hole, Rush and Beck, 1979)

Reviews:
Cited in Cochrane reviews (v1, v2 and v3) although it was excluded from the review because it is not an RCT.
Not cited in full NICE schizophrenia guidance 1985 Throughout the 1980s, consistent findings that family environments, in terms of the measure of expressed emotion (EE) were strongly influential in determining the recurrence of schizophrenic symptoms promoted the development of family interventions that were successful in reducing relapse.  (Tyrer et al, 1988(Tyrer et al, , 1990 in which cognitive therapy was employed by a group of patients with neurotic disorders. Kingdon realised that the techniques he was using with patients who had psychoses could be described as a variant of cognitive therapy. This prompted him and colleagues to define the techniques used, incorporate additional components, and refine them with further study of the literature. Development of cognitivebehavioural interventions for psychotic patients as direct therapies for specific symptoms (Chadwick and Lowe, 1990) and as a way of enhancing patients' coping skills (e.g. Tarrier et al., 1990) Human researchdosage/design (phase II) Publications: Chadwick, P. D. J. and Lowe, C. F. (1990) Kingdon and Turkington, 1991). N=64 Human researchdosage/design (phase II) Publications: Kingdon, D.G. and Turkington, D. (1991) 'The use of cognitive behaviour therapy with a normalising rationale in schizophrenia', Journal of Nervous and Mental Disease 179: 207-211.

Reviews:
Cited in Cochrane reviews (v1, v2 and v3) although excluded from the review because it is not an RCT.  Bentall et al., 1994;Chadwick and Birchwood, 1994;Garety et al., 1994) Narrative A key obstacle in developing the case study was identifying a range of calibration points such as proposal submitted, first patient recruited and dates of data collection. In most cases this information was not given in papers, and therefore where it was given it could not be used for reasons of consistency. If other case studies have used calibration points other than publication dates, the time lags between this work and others may not be comparable.
More generally, the matrix worked well in providing a visualisation of CBT's development. However, it did not provide any insight regarding the cause of time lags in CBT's development -particularly given that it does not capture negative events. If the matrix were to capture and clarify causes it would need to capture general narrative and context as well as just events -e.g. to show the lack of trained CBT practitioners in preventing the widespread adoption of CBT.

Time lags between conducting medical research and its translation. Case Study 7: Early intervention in schizophrenia NARRATIVE ACCOUNT OF BACKGROUND/ DEFINITIONS AND KEY ASPECTS OF THE INTERVENTION'S DEVELOPMENT
Early intervention in schizophrenia may refer to intervention in either the prodromal stage (i.e. to prevent onset of a psychotic episode) or the first episode stage (i.e. focusing on prompt detection and treatment of psychosis). Some studies have associated tailored first-episode interventions with an improvement in treatment response and long-term outcomes, and at the very least these engage people with care at an early stage in order to reduce suffering, but there have been few randomised controlled trials demonstrating a causal link. In the last ten years there have been efforts to reach individuals with early intervention and treatment in the prodromal stage. Though this is still an emerging area, there are some studies, including five randomised controlled clinical trials, which have demonstrated the potential effectiveness of prodromal interventions. Nonetheless, many argue that the evidence base is not considered strong and there remains debate in the literature and within the schizophrenia research community about the nature and robustness of the evidence base around the effectiveness of both prodromal interventions and first-episode interventions.
In the absence of national research and development strategies during the initial stages of development, many of the key events and initiatives in the development of early intervention services were driven at a local level -for example, the work of Patrick McGorry and Alison Yung at the EPPIC centre and PACE clinic in Australia. These intervention programmes and services for both the prodromal and first-episode stages not only established an initial evidence base, but also brought many researchers together to form an international network which began to establish similar programmes and services in other localities.
Despite locally strong initiatives in different countries, comparison between Canada, UK and USA suggests only the UK has evidence of a national research and clinical delivery strategy for intervention in the first-episode stage. There has also been considerable policy and advocacy activity at a national level, although this has triggered extensive debate, particularly around prodromal intervention. The suggestion that an ultra-high-risk group can be identified and treated based on prodromal symptoms is controversial, due to the possibility of 'false positive' diagnoses and unnecessary stigmatisation and discrimination. While a 'risk syndrome' was initially proposed for inclusion in DSM-V, this was subsequently dropped.
Due to the lack of clarity over the strength of the evidence base for the effectiveness of either form of early intervention, the practice guidelines in the USA, the UK and Canada vary in the degree to which they promote early intervention. Although the guidelines generally recognise the importance of providing treatment as early as possible and recommend intervention at the first-episode stage, they do not identify a specific tailored approach for this intervention, and none recommends prodromal interventions.  [1984][1985][1986][1987][1988] Buckingham study of early intervention in the prodromal stage (England) The first study to focus on early detection of psychosis in the prodromal period. not in scope ("Early intervention is primarily concerned with identification and initial treatment of people with psychotic illnesses, such as schizophrenia. Identification may be directed either at people in the prodromal phase of the illness ('earlier early intervention') or at those who have already developed psychosis ('early intervention'). Providing treatment for people in a possible prodromal phase of schizophrenia is an interesting but potentially controversial area, which at present is outside the scope of this guideline").    The approach the case study author took for this case study relied on tracing bibliographic references (citations) manually, but systematically, through the literature using the "Research review" calibration point as the key source for both forward and backward tracing. The key source for this was three versions of the Cochrane Collaboration review. This approach has advantages and disadvantages, especially when compared with other approaches different authors utilised in developing the other six case studies. Perhaps most importantly this approach has the potential to be automated through advanced bibliometric data mining techniques and potentially topic modelling (i.e. the examination of a corpus of literature over time using statistical methods). If such approaches could be developed then, crucially, the approach could be scaled at relatively modest costs. The disadvantage of this approach, however, is it relies exclusively on published (ie dated) publications that are in the public domain and accessible via web searching. One of the observations we made during the project was that the date of a publication may be misleading; it can occur a number of years -sometimes decades -after the research was completed or it can occur very closely to that research. Differential practice in the data of publication makes it difficult to interpret the data. That said publication and the publication date is an indicator of when the research is 'visible' and in that sense it may be an appropriate proxy for measuring elapsed time.
Another observation the case study author made in compiling this case study was the apparent 'connectedness' of the publishing authors -largely around Pat McGorry and colleagues. A hypothesis that could be explored through social network analysis is that the elapsed time between various stages was sufficiently speeded up due to the collaborative network that emanated from Australia.