This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Brealey, S Articles by Thomas, N B Search for Related Content PubMed PubMed Citation Articles by Brealey, S Articles by Thomas, N B

Methodological standards in radiographer plain film reading performance studies

¹Department of Health Sciences, University of York, York YO1 5DD, ²Division of Radiography, University of Bradford, Bradford BD5 0BB and ³X-ray Department A, North Manchester General Hospital, Manchester M8 5RB, UK

	Abstract

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
The objectives of this paper are to raise awareness of the methodological standards that can affect the quality of radiographer plain-film reading performance studies and to determine the frequency with which these standards are fulfilled. Multiple search methods identified 30 such studies from between 1971 and the end of June 1999. The percentage of studies that fulfilled criteria for the 10 methodological standards were as follows. (1) Performance of a sample size calculation, 3%; (2) definition of a normal and abnormal report, 97%; (3) description of the sequence of events through which films passed before reporting, 94%; (4) analysis of individual groups of observers within a combination of groups, 50% (5) appropriate choice of reference standard, 80%; (6) appropriate choice of arbiter, 57%; (7) appropriate use of a control, 22%; (8) analysis of pertinent clinical subgroups, e.g. body areas, patient type, 44%; (9) availability of data for re-calculation, 59%; and (10) presentation of indeterminate results, 69%. These findings indicate variation in the application of the methodological standards to studies of radiographer's film reading performance. Careful consideration of these standards is an essential component of study quality and hence the validity of the evidence base used to underpin radiographic reporting policy.

	Introduction

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
The NHS and Community Care Act (1990) produced a climate in which traditional methods of delivering health care were challenged, resulting in the blurring of professional boundaries [1]. Since then there has been an increase in the reporting of Accident and Emergency films by radiographers [2, 3].

When assessing radiographers' film reading performance it is not always possible, or necessary, to conduct a randomized controlled trial. It is therefore important to be aware of the main threats to study validity from the alternative designs encountered. Furthermore, there is a conceptual hurdle to overcome: how does one relate patient outcome to the reports made by different professionals when other factors, such as therapy, are involved? In the context of imaging technologies, this difficulty was resolved by applying a hierarchical framework first suggested by Fineberg et al [4] and subsequently extended by the Institute of Medicine [5]. The categories involved are technical capability, diagnostic accuracy, diagnostic impact, therapeutic impact and impact on health [6].

Film reading performance studies are comparable with the "diagnostic accuracy" category [7]. These studies involve observers e.g. radiographers, interpreting a sample of films. An arbiter, i.e. health care professional, then judges whether the reports made by the radiographers are concordant with a reference standard, e.g. consultant radiologist. The resulting data is then used to calculate performance statistics such as sensitivity and specificity. However, the environment in which a radiographer is assessed, i.e. controlled conditions or clinical practice can affect, for example, film selection, choice of reference standard and method of analysis [7]. Table 1 describes three different types of plain-film reading performance studies. Those that assess observers reporting in controlled conditions, such as radiographers under examination conditions, are called diagnostic accuracy studies. In such studies a mix of normal and abnormal films are carefully selected with abnormalities covering a range of pathology, body areas and degrees of conspicuity. A robust reference standard such as a double/triple blind consultant radiological report is developed against which the observers' reports are compared to ensure they interpret films to a high level of accuracy before reporting in clinical practice. Diagnostic performance studies monitor the progress of one group of observers interpreting a consecutive series of films during clinical practice compared, perhaps, with a single consultant radiologist as the reference standard. Studies that compare radiographers and other professional groups, such as casualty officers, against a reference standard are called diagnostic outcome studies because we assume that the group with the highest reporting accuracy will contribute more to improving patient outcome.

	Data sources

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
Four methods were used to identify eligible studies. These were electronic sources, hand-searching journals, personal communication and scanning reference lists. Eight electronic sources were searched; Medline (Index Medicus), Bath Information and Data Services, Cumulative Index to Nursing and Allied Health Literature, Embase (Excerpta Medica), NHS National Research Register, the Cochrane Library, PsycLIT (Psychological abstracts) and the System for Information on Grey Literature. The Medline search used a combination of terms derived from its thesaurus and from the terms used to index studies by known authors in the subject area. This included the single index terms "diagnostic-errors" and "sensitivity-and-specificity" and the exploded index terms "radiography" and "radiology". The text words used included "reporting", "radiographs", "radiographers", "triage" and "x ray film(s)". A similar strategy was developed for searching the other databases. The searches were performed, when possible, on literature published between 1971 and the end of June 1999. This was to coincide with the proposal in 1971 by Swinburne of radiographers being used to distinguish normal from abnormal films [12]. No language or geographical limits were incorporated into the search strategies.

The journals and supplements that were hand-searched included the British Journal of Radiology, Clinical Radiology, Radiography Today/Synergy and Radiography. With the exception of Radiography, for which issues from 1995 onwards were searched, journal searches covered issues between 1990 and the end of June 1999, as this was when the debate accelerated with the introduction of the NHS and Community Care Act [1]. The Royal College of Radiologists, the Society of Radiographers and the College of Radiographers were also contacted to identify studies, as were members of the Special Interest Group in Radiographic Reporting and university centres that provide postgraduate courses to train radiographers in film interpretation. The reference lists of all articles identified were scanned for further studies.

	Study selection

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
Studies were eligible for inclusion if they assessed the plain-film reading performance of one or more radiographers compared with a reference standard, and included appropriate statistics, e.g. sensitivity and specificity. Studies were excluded based on the following criteria: not performed during the specified time frame, non-English language, non-UK, insufficient data to apply selection criteria, case studies, visual search strategy studies, and abstracts later published as papers. Only original papers were used to minimize duplication of data when studies were re-published.

Searching electronic sources yielded 695 studies, of which 20 assessed radiographers' film reading performance. However, only seven of these studies were eligible. Of the 13 studies excluded, 1 was a visual search strategy study and the other 12 were the same studies identified in different databases. A total of 30 studies were judged eligible from all data sources.

To minimize "reviewer bias", studies from the electronic databases were selected independently, i.e. blindly, by two reviewers (SB and AS). Perfect agreement was found in the application of the selection criteria. It was therefore judged acceptable for only SB to apply these criteria to studies located from other data sources.

	Data extraction

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
All eligible studies were assessed independently by SB for the 10 methodological standards. When a standard was met it was scored as A, when not fulfilled as C and when not applicable as N/A. Data extraction could only be completed by personal communication for 10 of the studies. This involved SB visiting the investigator responsible or completing the checklist by telephone. For the remaining studies the other two reviewers (AS and NT) independently applied the standards for 10 studies each. All reviewers were familiar with many of the studies so no blinding was made of publication details. Any discordance between reviewers was resolved by discussion. SB, compared with AS and NT, respectively, had 89% and 86% agreement and substantial kappa scores of 0.82 (95% confidence interval (CI) 0.72–0.92) and 0.79 (95% CI 0.69–0.89).

	Data synthesis

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
Provided below is a description of each of the methodological standards along with the criteria used to judge whether that standard had been fulfilled and the frequency with which the standard was met. Table 2 lists the questions used to assess whether the standards were met. The results in Table 3 show the number and percentage of standards met when applicable for the three different study types presented in Table 1. Table 4 shows the number and percentage of all the plain-film reading performance studies that met the methodological standards when applicable.

The standard criteria used were (a) a study that measured the performance of a single group of radiographers had to calculate a sample size according to how precise an estimate of sensitivity and specificity was needed and (b) a study comparing groups had to use a power calculation to determine the sample size required to detect clinically important effects as statistically significant.

Table 3 shows none of the 11 diagnostic accuracy studies calculated the necessary sample size. Six studies used a sample of fewer than 100 films. Only one diagnostic performance study calculated the necessary sample size. However, the number of films included in these studies ranged from several hundred to several thousand. No diagnostic outcome study performed a sample size calculation, although each study used a sample of films ranging around several hundred. Thus, Table 4 shows that only 1 (3%) of the radiographer film reading performance studies satisfied this standard.

Study design
Standard 2: was a normal/abnormal report adequately defined?
The definition of what is "normal" or "abnormal" influences both sensitivity and specificity. These individual measures are therefore subjective in that an investigator can define what they are, although a pair of sensitivity/specificity values are not subjective [8]. If studies use different criteria to define positive and negative reports this will affect the validity of comparisons between them. This is because differences in performance statistics may be entirely due to variation in these definitions. Subsequently, any published set of estimates is of limited value unless what constituted a normal or abnormal report is adequately defined.

Only one standard criterion was used: whether the definition of normal or abnormal is acceptable depended on the context of the study.

Apart from one diagnostic performance study, all studies adequately described what was normal and abnormal. Indeed, Table 4 shows that this standard was met in 28 (97%) studies.

Standard 3: was the performance of the observers placed in the context of the diagnostic sequence?
For studies conducted in clinical practice it is important to describe the process, or diagnostic sequence, through which films pass. The point at which radiographers interpret the films will affect the sample in terms of prevalence and severity of disease. This will affect the predictive values and possibly sensitivity and specificity. Without this information readers will not be able to apply the results of the study to the system employed in their hospital.

Standard criterion: the process through which the films passed before interpretation must be explicitly described. This standard was not applicable to studies performed outside of clinical practice, i.e. diagnostic accuracy studies.

Again, except for one diagnostic performance study, all of the 17 (94%) applicable studies adequately described the context in which observers reported.

Standard 4: was the contribution of individual groups determined if the combined performance of two (or more) different groups of observers was assessed?
If the aim of a study is to assess the combined performance of two groups, e.g. radiographers and emergency nurse practitioners, it is desirable to assess the individual contribution of each group. This will demonstrate the marginal utility of the combined groups compared with each individual group.

Standard criterion: this standard was met if each group within a combination of groups was assessed independently.

Table 3 demonstrates that this standard was not applicable to diagnostic accuracy or performance studies as no studies assessed the combined film reading performance of two or more groups of observers. 1 (50%) of 2 diagnostic outcome studies met this standard.

Standard 5: was an appropriate (valid) reference standard used?
The validity of radiographers' film reading performance is dependent on the veracity of the reference standard [16]. Considerable variation has been found between experienced radiologists when interpreting plain-films [17]. Therefore, convergence of multiple radiologists' opinions should provide a better reference standard than one radiologist [16]. Searching for further examinations of the same body area over a follow-up period of 1 year could be used to validate the accuracy of a single radiologist's report. Findings at follow-up could indicate whether the reference standard report was erroneous [18].

The criteria for this standard involved using the following hierarchy to judge the validity of the reference standard, with A1 being the most valid. (A1) A double/triple blind consultant radiological report; (A2) a single blind consultant radiological report validated using, for example, clinical follow-up; and (A3) a single blind consultant radiological report. The standard was not fulfilled if an inappropriate reference standard was used, e.g. a combination of radiologists at different grades, or if the professionals under evaluation were used as the reference standard or included in the process of generating the reference standard.

All 11 (100%) diagnostic accuracy studies used a valid reference standard; 6 used a double/triple blind consultant radiological report (A1). In contrast, 5 (45%) diagnostic performance studies failed to meet this standard. Four of these failed because they included the reports of radiographers under evaluation in the process of developing the reference standard, and the fifth study failed because it used junior radiologists' reports. Furthermore, the most frequently used valid reference standard was only a single consultant radiological report (A3). For diagnostic outcome studies all reference standards chosen were valid with the exception of one study that used a consultant radiologist or specialist registrar. Table 4 shows that the appropriate use of a valid standard was met in 24 (80%) of all studies.

Standard 6: was an appropriate (valid) arbiter used to compare radiographers' reports with the reference standard?
Critical elements of film interpretation are knowing what to look for in images and why [19]. Similarly, the professional, or arbiter, responsible for comparing reports should possess this knowledge. The primary criterion used to judge the validity of the arbiter is whether they were external to the institution under evaluation, i.e. where the arbiter was based. An external arbiter should be more objective than an internal one, as their institution is not being evaluated. Furthermore, even if an internal arbiter is blind to which report was made by whom, they might recognize who made a report which could consciously or unconsciously affect their judgement. The second criterion is whether the study used a panel of arbiters rather than an individual, i.e. the number of arbiters used. Even when explicit criteria are available for comparing reports there can be variation in how they are applied by different arbiters. This is analogous with the discussion of single vs multiple radiologists producing a reference standard report. Therefore when a panel of arbiters were involved in the process of comparing reports this was judged as being more valid than a single arbiter. The final criterion focused on whether the arbiter was appropriately skilled to perform this task, i.e. who the arbiter was.

Standard criteria: the following hierarchy was used to judge the validity of the arbiters with A1 being the most valid. (A1) External panel; (A2) external single consultant radiologist; (A3) internal panel; (A4) internal single consultant radiologist; (A5) radiographer trained to report supported by an independent consultant radiologist; and (A6) untrained radiographer(s) supported by an independent consultant radiologist. Examples of inappropriate arbiters include untrained radiographer(s) with no referral to a radiologist in equivocal cases, and a professional under evaluation.

9 (82%) of the 11 diagnostic accuracy studies used a valid arbiter, with 3 (27%) studies using an internal panel (A3). However, radiographers not trained to report were used in one study with no option to refer to a radiologist, and in one other study it was not possible to discern who was the arbiter. In contrast, the use of a valid arbiter was satisfied in only 2 (18%) of the 11 diagnostic performance studies. 7 (64%) of these studies used the radiographer under evaluation as the arbiter, or the radiographer was included in the process of arbitration. It was unclear who the arbiter was in two studies. Finally, 6 (75%) of the 8 diagnostic outcome studies used a valid arbiter. Of the two studies that did not, one used a radiographer not trained to report and the other study did not refer to whom the arbiter was. Table 4 shows that the appropriate use of a valid arbiter was met in 17 (57%) of studies.

Standard 7: was an appropriate control used?
Controls help rule out potential threats to study validity and eliminate the possibility of alternative explanations [20]. To assess the effectiveness of a training programme (intervention) on radiographers' film reading performance there should be a group who received intervention (experimental) and a group who did not (control). The two groups should be matched for appropriate characteristics, such as number of years experience in the profession or in a relevant speciality [21]. Improved performance in the experimental group could then be attributed to the intervention rather than differences in the sample of films or radiographers.

Standard criterion: this standard was met if an appropriate control was used within the context of the study.

When applicable, 2 (50%) of 4 diagnostic accuracy studies did not use a control group when assessing the effectiveness of a training course on radiographers' ability to interpret films. None of the three diagnostic performance studies used a control. Two assessed the effectiveness of a training programme and another the effect of introducing a radiographer abnormality detection system (red dot system) on casualty officers' error rates. Similarly, neither of the outcome studies met this standard. Table 4 shows that only 2 (22%) studies used an appropriate control.

Presentation of results
Standard 8: were films appropriately analysed for pertinent subgroups?
Even if the case mix of films has been adequately described, the performance statistics represent average values for the entire sample that may mask low levels of accuracy for a particular subgroup [11].

Standard criterion: the standard was met if radiographers' performance was presented for pertinent medical subgroups, e.g. body areas, patient type.

Table 3 shows that in general, for diagnostic accuracy, performance and outcome studies, around 50% met this standard. Table 4 shows that it was met in 12 (44%) of all studies.

Standard 9: were the data presented in enough detail to permit re-calculation of performance statistics and confidence intervals?
The presentation of raw data is necessary so that readers or reviewers can calculate relevant performance statistics and confidence intervals. Confidence intervals are important to illustrate the range of values we can be confident includes a radiographer's true film reading performance [15]. The width of the interval indicates whether the sample size is too small to draw a valid conclusion from.

Standard criterion: data needed to be presented in enough detail to permit the calculation of performance statistics and confidence intervals.

Table 3 shows that 8 (73%) and 5 (63%), respectively, diagnostic performance and outcome studies achieved this standard. Overall this standard was met in 16 (59%) studies.

Standard 10: were indeterminate, i.e. equivocal or missing data, results appropriately presented?
In clinical practice it is not always possible to provide a clear-cut interpretation of a film owing to factors such as technical defects and artefacts, patient restrictions or administrative limitations [22]. Performance statistics are therefore distorted if there is no option to classify a report as being equivocal [23]. If indeterminate results are included but regarded as positive, sensitivity is artificially increased and specificity decreased. The reverse effects occur if the indeterminate results are counted as negative. Corresponding distortions occur in the calculation of likelihood ratios [11].

Two standard criteria had to be met before this standard was fulfilled. First, the study must present all of the appropriate positive, negative and indeterminate interpretations. Second, the study must describe whether indeterminate interpretations had been included or excluded when performance statistics were calculated.

6 (75%) and 3 (75%), respectively, diagnostic performance and outcome studies met this standard, compared with 2 (50%) diagnostic accuracy studies. A possible explanation for this is that because diagnostic accuracy studies use a carefully selected sample of films to assess radiographers' film reading performance, unequivocal cases were mostly used. Table 4 shows this standard was met in 11 (69%) studies overall.

	Conclusions

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 
A combination of biases and methodological factors influence the validity of film reading performance studies. A recent paper describes these biases [24] and a review of the prevalence of bias in these studies is the subject of a further paper (not yet published). The results of this review demonstrate wide variation in the fulfilment of the 10 methodological standards, such as the use of a sample size calculation (3%), which affect precision of findings, and the choice of reference standard (80%) and arbiter (50%), which can distort film reading performance statistics.

Ideally, all standards should be met. However, possibly owing to resource constraints and study objectives, there are trade-offs between standards. As an example, when a robust reference standard is required, as in diagnostic accuracy studies, fewer films are selected. Conversely, when a large sample of films is required, as in diagnostic performance studies, a less robust reference standard is used. The challenge to those designing and executing such studies is to achieve the right balance between the methodological standards. If these standards are not adhered to it will increase the chance of erroneous conclusions being made about radiographers' film reading performance. This in turn can affect radiographic reporting policy and, ultimately, patient care and service efficiency.

	Acknowledgments

 
We are profoundly grateful for the assistance of the authors of the studies included in this review, without whose help this paper would not be possible. We also thank the two referees for their comments.

Received for publication April 17, 2001. Accepted for publication November 8, 2001.

	References

Top Abstract Introduction Data sources Study selection Data extraction Data synthesis Conclusions References

 

1. The College of Radiographers. Role development in radiography. London: College of Radiographers 1996.
2. Paterson AM. Role development—towards 2000: a survey of role developments in radiography. London: College of Radiographers, 1995.
3. Price RC, Le Masurier SB, High L, Miller LR. Changing times: a national survey of extended roles in diagnostic radiography. Br J Radiol 1999; 72(Suppl.):7.
4. Fineberg HV, Bauman R, Sosman M. Computerized cranial tomography: effect on diagnostic and therapeutic plans. JAMA 1977;238:224–7.[Abstract/Free Full Text]
5. Institute of Medicine. Policy statement: computed tomographic scanning. Washington DC: National Academy of Sciences, 1977.
6. Mackenzie R, Dixon AK. Measuring the effects of imaging: an evaluative framework. Clin Radiol 1995;50:513–8.[Medline]
7. Brealey S. Measuring the effect of image interpretation: an evaluative framework. Clin Radiol 2001;56:341–7.[Medline]
8. Begg CB. Biases in the assessment of diagnostic tests. Stat Med 1987;6:411–23.[Medline]
9. Irwig L, Tosteson ANA, Gatsonis C, Lau J, Colditz G, Chalmers TC, et al. Guidelines for Meta-analyses evaluating diagnostic tests. Ann Intern Med 1994;120:667–76.[Abstract/Free Full Text]
10. Jaeschke R, Guyatt G, Sackett DL. Users' guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid? Evidence Based Medicine Working Group. JAMA 1994;271:389–91.[Abstract/Free Full Text]
11. Carrington Reid RM, Lachs MS, Feinstein AR. Use of methodological standards in diagnostic test research: getting better but still not good. JAMA 1995;274:645–51.[Abstract/Free Full Text]
12. Swineburne K. Pattern recognition for radiographers. Lancet 1971;I:589–90.
13. Freedman LS. Evaluating and comparing imaging techniques: a review and classification of study designs. Br J Radiol 1987;60:1071–81.[Abstract/Free Full Text]
14. Arkin CF, Wachtel MD. How many patients are necessary to assess test performance? JAMA 1990;263:275–8.[Abstract/Free Full Text]
15. Altman DG. Clinical Trials. In: Practical statistics for medical research. London: Chapman & Hall, 1991:440–74.
16. Robinson PJA. Radiology's Achilles' heel: error and variation in the interpretation of the Rontgen image. Br J Radiol 1997;70:1085–98.[Abstract]
17. Robinson PJA, Wilson D, Coral A, Murphy A, Verow P. Variation between experienced observers in the interpretation of accident and emergency radiographs. Br J Radiol 1999;72:323–30.[Abstract]
18. Robinson PJ. Plain film reporting by radiographers—a feasibility study. Br J Radiol 1996;69:1171–4.[Abstract/Free Full Text]
19. Robinson PJA. The nature of image reporting. In: Paterson A, Price R, editors. Current topics in radiography. London: WB Saunders, 1996:70–82.
20. Cook TD, Campbell DT. Causal inference and the language of experimentation. In: Quasi-Experimentation—Design & Analysis Issues for Field Settings. London: Houghton Mifflin Company, 1979:1–36.
21. Maisey MN, Hutton J. Errors associated with measuring diagnostic performances. In: Guidelines for the evaluation of radiological technologies—report of a working party of the British Institute of Radiology. London: British Institute of Radiology 1991:18–9.
22. Robinson PJA, Fletcher JM. Clinical coding in radiology. Imaging 1994;6:133–42.
23. Simel DL, Feussner JR, Delong ER, Matchar DB. Intermediate, indeterminate, and uninterpretable diagnostic test results. Medic Decis Making 1987;7:107–14.
24. Brealey S, Scally AJ. Bias in plain film reading performance studies. Br J Radiol 2001;74:307–16.[Abstract/Free Full Text]

This article has been cited by other articles:

				S BREALEY, C HEWITT, A SCALLY, S HAHN, C GODFREY, and N THOMAS Bivariate meta-analysis of sensitivity and specificity of radiographers' plain radiograph reporting in clinical practice Br. J. Radiol., July 1, 2009; 82(979): 600 - 604. [Abstract] [Full Text] [PDF]

				S D Brealey, A J Scally, S Hahn, and C Godfrey Evidence of reference standard related bias in studies of plain radiograph reading performance: a meta-regression Br. J. Radiol., June 1, 2007; 80(954): 406 - 413. [Abstract] [Full Text] [PDF]

				S Brealey, D G King, S Hahn, C Godfrey, M T I Crowe, K Bloor, S Crane, and D Longsworth The costs and effects of introducing selectively trained radiographers to an A&E reporting service: a retrospective controlled before and after study Br. J. Radiol., June 1, 2005; 78(930): 499 - 505. [Abstract] [Full Text] [PDF]

				P. M. Bossuyt, J. B. Reitsma, D. E. Bruns, C. A. Gatsonis, P. P. Glasziou, L. M. Irwig, D. Moher, D. Rennie, H. C.W. de Vet, and J. G. Lijmer The STARD Statement for Reporting Studies of Diagnostic Accuracy: Explanation and Elaboration Ann Intern Med, January 7, 2003; 138(1): W1 - W12. [Abstract] [Full Text] [PDF]

				P. M. Bossuyt, J. B. Reitsma, D. E. Bruns, C. A. Gatsonis, P. P. Glasziou, L. M. Irwig, D. Moher, D. Rennie, H. C.W. de Vet, and J. G. Lijmer The STARD Statement for Reporting Studies of Diagnostic Accuracy: Explanation and Elaboration Clin. Chem., January 1, 2003; 49(1): 7 - 18. [Abstract] [Full Text] [PDF]

				S Brealey, D G King, M T I Crowe, I Crawshaw, L Ford, N G Warnock, R A J Mannion, and S Ethell Accident and Emergency and General Practitioner plain radiograph reporting by radiographers and radiologists: a quasi-randomized controlled trial Br. J. Radiol., January 1, 2003; 76(901): 57 - 61. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Brealey, S Articles by Thomas, N B Search for Related Content PubMed PubMed Citation Articles by Brealey, S Articles by Thomas, N B