NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Pandor A, Goodacre S, Harnan S, et al. Diagnostic Management Strategies for Adults and Children with Minor Head Injury: A Systematic Review and an Economic Evaluation. Southampton (UK): NIHR Journals Library; 2011 Aug. (Health Technology Assessment, No. 15.27.)

Cover of Diagnostic Management Strategies for Adults and Children with Minor Head Injury: A Systematic Review and an Economic Evaluation

Diagnostic Management Strategies for Adults and Children with Minor Head Injury: A Systematic Review and an Economic Evaluation.

Show details

Appendix 13Protocol

Project title

The cost-effectiveness of investigation and hospital admission for minor (Glasgow Coma Scale 13–15) head injury.

Planned investigation

Research objectives

We aim to identify the optimal strategy for managing adults and children with minor [Glasgow Coma Scale (GCS) 13–15] head injury. Our specific objectives are to:

  • estimate the diagnostic accuracy of clinical assessment, clinical decision rules, skull radiography, cranial computerised tomography (CT) and inpatient observation for identifying intracranial bleeding requiring neurosurgery in adults and children with minor (GCS 13–15) head injury
  • estimate the cost-effectiveness of diagnostic strategies for minor head injury (MHI), in terms of the cost per quality-adjusted life-year (QALY) gained by each strategy
  • identify the optimal strategy for managing MHI in the NHS, defined as the most cost-effective strategy at the National Institute for Health and Clinical Excellence (NICE) threshold for willingness to pay per QALY gained
  • identify the critical areas of uncertainty in the management of MHI, where future primary research would produce the most benefit.

Existing research

Head injury is responsible for around 700,000 emergency department (ED) attendances per year in England and Wales, most of which (90%) will be minor (GCS 13–15) and will not need immediate neurosurgical intervention or inpatient care.1 These patients have a small (< 1%), but important risk of subsequent deterioration due to intracranial bleeding. If these cases are recognised and treated early then a full recovery can be expected; if not then severe disability or death may ensue.

Potential diagnostic management strategies for MHI typically use a combination of clinical assessment, skull radiography, CT scanning and hospital admission for observation to detect intracranial bleeding. The choice of strategy will have substantial cost implications for the NHS because it will be applied to hundreds of thousands of patients each year. Only a small proportion of patients will have intracranial bleeding, but those who do have a huge potential to benefit from early diagnosis and treatment.

Guidelines for managing head injury were drawn up by NICE in 20031 and revised in 2007.2 These guidelines were based upon literature review and expert consensus. Cost-effectiveness analysis was not used to develop the guidelines, but was used to explore the potential impact upon health service costs. The guidelines were expected to potentially reduce costs, but recent data (outlined below) suggest that costs may have substantially increased. Additional expenditure on MHI may represent a worthwhile use of NHS resources but there are currently no relevant analyses to support this. An extensive evidence synthesis and economic evaluation is thus needed to inform future NICE guidance.

Clinical assessment can be used to identify patients with an increased risk of intracranial bleeding and select patients for imaging or admission. A recent meta-analysis of 35 studies reporting data from 83,636 adults with head injury3 found that severe headache (relative risk 2.44), nausea (2.16), vomiting (2.13), loss of consciousness (LOC) (2.29), amnesia (1.32), post-traumatic seizure (PTS) (3.24), old age (3.70), male gender (1.26), fall from a height (1.61), pedestrian crash victim (1.70), abnormal GCS score (5.58), focal neurology (1.80) and evidence of alcohol intake (1.62) were all associated with intracranial bleeding. A similar analysis of 16 studies reporting data from 22,420 children with head injury4 found that focal neurology (9.43), LOC (2.23) and abnormal GCS score (5.51) were associated with intracranial bleeding.

Clinical features have been combined in a number of studies to develop a structured clinical decision rule. A systematic review undertaken for the NICE guidance1 identified four studies of four different clinical decision rules. The studies of the Canadian CT rule5 and the New Orleans rule6 were both high quality, applicable to the NHS and reported 100% sensitivity for the need for neurosurgical intervention. The other two studies,7,8 respectively, reported poor sensitivity and were not applicable to the NHS.

Several further studies have been published and new rules developed since the NICE review. A comparison of the Canadian CT and New Orleans rules undertaken by the researchers who developed the Canadian rule9 showed that both rules had 100% sensitivity for predicting neurosurgical intervention and clinically important brain injury, but the Canadian rule had higher specificity (76.3% vs 12.1% and 50.6% vs 12.7%). A comparison by an independent team10 found that both rules had 100% sensitivity for neurosurgical intervention, but the New Orleans rule had higher sensitivity for clinically significant brain injury (99.4% vs 87.2%), while the Canadian rule had higher specificity (39.7% vs 5.6%). This team also developed a new rule, the CT in Head Injury Patients (CHIP) rule, with 100% sensitivity and 30% specificity for neurosurgical intervention.11

New rules have also been developed for children with head injury. The National Emergency X-Radiography Utilization Study II (NEXUS II) rule, was developed and shown to have 98.6% sensitivity and 15.1% specificity for significant ICI,12 whereas the Children's Head injury Algorithm for the prediction of Important Clinical Events (CHALICE) rule had 98.6% sensitivity and 86.9% specificity.13 The striking difference in specificity may be due to the use of broader selection criteria in the CHALICE study (and thus lower prevalence). Both studies had similar positive predictive value (9.5% vs 8.6%) and negative predictive value (99.1% vs 99.9%).

Skull radiography can identify fractures that are associated with a substantially increased risk of intracranial bleeding, but cannot identify intracranial bleeding itself. Skull radiography is therefore used as a screening tool to select patient for investigation or admission, but not for definitive imaging. A meta-analysis14 found that skull fracture detected on radiograph had a sensitivity of 38% and specificity of 95% for intracranial bleeding. More recent meta-analyses in adults3 and children4 reported relative risks of 4.08 and 6.13, respectively, for the association between skull fracture and intracranial bleeding.

Computerised tomography definitively shows significant bleeding and a normal CT scan effectively excludes a significant bleed at the time of scanning. MRI scanning can detect some lesions that are not evident on CT,15 but arguably none that is of clinical importance and certainly none that influences early management. CT can therefore be considered a reference standard investigation for detecting injuries of immediate clinical importance.

Hospital admission and observation may be used to identify intracranial bleeding by monitoring the patient for neurological deterioration. Although commonly used in the past, the effectiveness of this approach has not been studied extensively and has the disadvantage that neurosurgical intervention is delayed until after patient deterioration has occurred. Hospital admission and observation are usually used selectively, based upon clinical assessment or skull radiography findings.

Theoretically, patients without intracranial bleeding on their CT scan do not require hospital admission. In practice, however, patients may be admitted for a number of reasons (1) CT scanning may identify abnormalities, such as minor cerebral contusions, which do not require neurosurgery and are of uncertain significance, but prompt hospital admission; (2) patients may be admitted pending CT scanning because they are deemed to need imaging but are unable to have imaging, either due to lack of availability or lack of ability to cooperate; and (3) patients may be admitted despite a normal CT because of concern about continuing symptoms, such as severe headache or vomiting, or with drug or alcohol intoxication.

Studies have compared CT-based strategies to skull radiography and/or admission to conclude that CT-based strategies are more likely to detect intracranial bleeding and less likely to require hospital admission.16,17 Cost analyses based upon trial data18 and modelling19 both suggest that a CT-based strategy is cheaper. However, admission-based strategies may be an inappropriate comparator for cost-effectiveness analysis because they appear to be expensive and ineffective, particularly if applied unselectively.

Computerised tomography may be used unselectively (in all patients) or selectively, based upon clinical assessment or a decision rule. A strategy of CT scanning all patients would clearly be very effective, but would have a low yield of positive results and would be expensive. The more selective the use of investigations or admission the cheaper the strategy, but the higher the risk of missed pathology. Cost-effectiveness analysis is therefore necessary to determine what level of investigation represents the most efficient use of health-care resources.

A study from the USA20 used decision analysis modelling to examine the cost-effectiveness of strategies for managing MHI and concluded that strategies involving selective CT use or CT for all, followed by discharge if negative, were cost-effective, whereas admission-based strategies were not. There was only limited exploration of uncertainty, particularly around the estimate of the effect of early versus delayed neurosurgery, and it is not clear whether the results are applicable to the NHS.

Despite the economic importance of MHI there has been little evaluation of cost-effectiveness in the NHS. Recent NICE head injury guidance was based upon the Canadian CT head rule9 and was anticipated to lead to more CT scans being performed, but fewer skull radiographs and admissions. A cost analysis1 suggested that the guidelines would be cost saving, by virtue of decreasing skull radiography and admissions while increasing CT scanning. Patient outcomes were not examined and the discussion cautioned that the assumption that increased CT scanning would reduce admissions might not hold in practice.

Data from a number of studies have since confirmed that more CT scans are being performed and less skull radiography is being undertaken.2123 However, Hospital Episode Statistics (HES) for England show that the annual number of admissions for head injury has increased from 114,769 in 2001–2 to 155,996 in 2006–7. As average length of stay has remained relatively constant, bed-days have increased from 348,032 in 2001–2 to 443,593 in 2006–7. As Figure 1 shows, the increase in admissions has been seen in adults rather than children.24

FIGURE 1. Head injury admissions in England, 1998–2007.

FIGURE 1

Head injury admissions in England, 1998–2007.

These data suggest that the annual costs of admission for head injury have increased from around £170M to £213M since the guidelines were introduced. Additional expenditure may be justified if associated with improved outcomes, but the anticipated effect of the guidelines was originally estimated only in resource terms and published studies have not examined effects upon patient health. It is therefore not clear whether this additional expenditure has produced any health benefits.

Management guidelines of the NHS should be based upon rigorous cost-effectiveness analysis. This is particularly important for MHI, where guideline development involves a trade-off between the costs of investigation and the benefits of detecting pathology, and where guideline implementation has substantial resource implications for the NHS.

Research methods

Design

We plan to undertake a cost-effectiveness analysis based on secondary research (systematic review, meta-analysis and decision-analysis modelling), along with a national survey and analysis of routine data sources to determine the most appropriate diagnostic management strategy for adults and children with minor head injuries in the NHS.

Systematic review and meta-analysis

Using standard methodology, we will undertake systematic literature reviews to identify:

  • cohort studies of patients with head injury that measure the diagnostic accuracy of any element of clinical assessment, any clinical decision rule, skull radiography, cranial CT or observation strategy for identifying intracranial injuries that require neurosurgery
  • observational or experimental studies that evaluate diagnostic management strategies for MHI in terms of process measures (hospital admissions, length of stay, time to neurosurgery) or patient outcomes
  • studies that report data to estimate key parameters in the decision-analysis model: prevalence of intracranial bleeding in MHI, survival and QoL after early or delayed neurosurgery for intracranial bleeding and long-term costs of care after neurosurgery for intracranial bleeding.

Search strategy

Relevant studies will be identified through electronic searches of key databases including MEDLINE, EMBASE, Science Citation Index (SCI) and Biological Abstracts. Recent published empirical work will be used to identify optimal strategies for prognosis and diagnosis on MEDLINE and EMBASE.2528

Search terms will include:

  • head injur$, craniocerebral trauma (including brain injuries, coma, post-head injur, cranial nerve injuries, head injuries (closed), brain concussion, head injuries (penetrating), intracranial haemorrhage (traumatic) and skull fracture)
  • clinical assessment, clinical decision rule$, guideline$, Canadian CT, CHIP, NEXUS, New Orleans, skull radiograph$, skull X-ray$, CT scanning, and hospital admission; plus such terms as
  • cohort studies, longitudinal studies, follow-up studies, time factors, long term, sequela$, prognosis
  • diagnostic terms such as specificity and sensitivity, false positive$, false negative$, true positive$, true negative$.

References will also be located through review of reference lists for relevant articles and through use of citation search facilities through WoK's SCI and Social Science Citation Index. Where existing systematic reviews already exist, these will be used both to identify relevant studies and to inform subsequent analysis. In addition, systematic searches of the internet the Copernic meta-search engine will be used to identify unpublished materials and work in progress. Key authors and professional and academic research groups will also be contacted and asked for unpublished material.

Review strategy

The stages of the review for diagnostic cohort studies will include:

  • Accumulation of references, entry and tagging on a Reference Manager database, enabling studies to be retrieved in each of the above categories by either keyword or textword searches.
  • Two reviewers will independently undertake preliminary review to identify any potentially relevant article based on titles, abstracts and subject indexing. All studies identified for inclusion, together with those for which a decision on inclusion is not possible from these brief details, will be obtained for more detailed appraisal.
  • Two reviewers will make decisions on the final composition of included studies, assessed from a hard copy of the item. The decisions will be coded and recorded on the Reference Manager database by the project manager.
  • Authors will be contacted, if appropriate, to clarify details and obtain missing data.
  • The quality of each study will be assessed against recognised criteria.29,30
  • Data extraction will be undertaken independently with discrepancies being discussed by the data extractors. Those that cannot be resolved at this stage will be referred to the rest of the project team.

These methods will also be used to identify studies of the management of head injuries and studies reporting data to inform the decision analysis model, but search terms, filters, selection criteria and quality assessments will be adapted to suit the purpose of each literature search.

Data extraction

The following data will be extracted from each study: population characteristics (age, gender, mechanism of injury, median GCS), setting (ED, general ward, neurosurgical centre), characteristics of the assessment or intervention (e.g. method of recording clinical features or decision score, staff training), definition of each outcome used, methods used to measure outcomes, study quality criteria (independence of the reference standard, blinding of the intervention and reference standard), prevalence of each outcome (clinically significant brain injury and need for neurosurgery), and true-positives, false-positives, false-negatives and true-positives for each outcome.

Data synthesis

Where appropriate, we will combine diagnostic data to provide pooled estimates of the diagnostic accuracy of clinical characteristics or clinical decision rules for diagnosing intracranial bleeding. For each modality, we will estimate the diagnostic performance (together with associated uncertainty) for diagnosing (1) intracranial bleeding requiring neurosurgery and (2) any clinically significant brain injury.

We will analyse data from adults and children separately wherever possible. Although we are specifically interested in diagnostic performance in patients with MHI we anticipate that most studies will report cohorts that include a range of severity. We will explore the applicability of findings to patients with MHI as part of our analysis of heterogeneity (see below).

The model used to analyse the data will depend on characteristics of the data obtained. For example, if diagnostic thresholds can be assumed constant across studies then simple methods of pooling sensitivity and specificity will be conducted.31 If there is implicit or explicit evidence that diagnostic thresholds differ between primary studies then sensitivity and specificity cannot be considered independent and simultaneous modelling will be required.32 A detailed assessment of heterogeneity will be conducted in all instances. If possible, meta-regression will be used to explore whether heterogeneity can be explained by study population characteristics, the method of implementation of the intervention, the definition of the outcome or the study quality, although the feasibility of this will depend on the number of individual studies identified and the quality of reporting. Where exploration of covariates is not possible or (unexplained) heterogeneity remains after the incorporation of covariates into the model(s), random effects will be incorporated to allow for such variability in results between studies.

Covariate effects, unexplainable variability and uncertainty in parameter estimates will all be reflected in the results using cutting-edge meta-analysis approaches. As the outputs from these analyses will be used in the decision modelling, all such sources of variation and uncertainty will be accurately reflected in the decision modelling.33

Standard meta-analysis methods will be used to combine multiple estimates, where they exist, for other parameters in the decision model.

A combination of Stata and the Meta-Disc statistical software34 (version Beta 1.0.10) will be used for this analysis.

Identification of potential management strategies

We will identify potential management strategies for MHI using the following methods:

  • Literature review As outlined above, we will identify any diagnostic management strategies evaluated in previous studies, particularly those based upon clinical decision rules.
  • Expert panel review We will constitute an expert panel of emergency physicians, neurosurgeons and neuroradiologists, who will review emerging data from the systematic reviews and then use consensus methods to develop potential diagnostic management strategies that would be appropriate for the NHS. These may be based upon established strategies or clinical decision rules, or theoretical combinations of clinical features and diagnostic tests identified as being diagnostically useful in the systematic reviews.
  • National survey We will undertake a national survey, as outlined below, to identify diagnostic management strategies that are currently being used in the NHS. These will then be reviewed by the expert panel and consensus methods used to select those with the potential for widespread use throughout the NHS.

National survey and routine data sources

We will undertake a national survey of EDs to identify formal guidelines used for MHI, clinical assessment strategies, policies for access to skull radiography and cranial CT, hospital admission policies (e.g. clinical decision unit, A&E observation or formal admission), bed availability, specialty responsible for inpatient care, staffing and senior supervision. This will be correlated with data from routine sources (e.g. HES).

We used a national survey in this way in our previous National Coordinating Centre for Health Technology Assessment (NCCHTA)-funded secondary research on diagnostic tests for deep vein thrombosis35 and found it to be a valuable source of data, and well worth the relatively trivial outlay of resources required to undertake it. Data from the national survey will provide the following:

  • identification of potential management strategies that are feasible in the NHS and can be evaluated by the decision-analysis model
  • data to inform the structure and populate key parameters of the decision-analysis model
  • context for our analysis, thus ensuring that the output of our research is relevant to the NHS.

Decision-analysis modelling

We will develop a decision-analysis model to estimate the costs and QALYs accrued by each potential management strategy for MHI, including a theoretical ‘zero option’ strategy of discharging all patients home without investigation. Each strategy will be applied to a theoretical cohort of patients attending the ED, with MHI allowing a direct comparison of results. For each strategy, sensitivity and specificity estimates from the literature review will determine the proportion of patients with intracranial bleeding who receive early or delayed neurosurgery and the proportion with no neurosurgical lesion who undergo diagnostic testing and/or admission to hospital.

The following costs will be estimated using data from the literature review, national survey, routine data sources and, if necessary, an expert panel: initial assessment, diagnostic tests (CT and skull radiography), hospital admission, neurosurgical intervention, long-term health and social care, and productivity losses.

Outcomes will be estimated as QALYs accrued following the decision to employ each management strategy. The expected utility associated with early or delayed neurosurgery will be taken from previous studies or, if necessary, expert panel opinion. We will search the literature to identify studies reporting survival and quality of life (QoL) after uncomplicated MHI (no bleeding), intracranial bleeding with early surgery, intracranial bleeding with delayed surgery and the disutility of the surgical procedure.

We will also use data from the Health And Long term Outcomes (HALO) study of patients with trauma. Researchers at the Medical Care Research Unit have been collecting diagnosis and baseline GCS, along with costs and QoL data up to 15 years after significant injury (including head injury). Where data from the existing literature are limited or inadequate we will ask the expert panel to review potential alternative data sources, for example extrapolating QoL data from other disabling neurological conditions. We will also use expert panel input to ensure that parameters are used in the model with appropriate estimates of uncertainty.

The time frame for the model will be the lifetime of the patient. We will assume that only patients with intracranial bleeding will incur long-term costs that are likely to be influenced by their initial diagnostic management, so long-term costs will be estimated only for patients in the model who survive intracranial bleeding. We will estimate discounted long-term costs by extrapolating follow-up costs from patients with significant head injury to the HALO study over the anticipated lifetime of the patient. Sensitivity analysis will be used to explore uncertainty in estimates of long-term costs. The baseline analysis will not include productivity losses but secondary analysis will be undertaken, including productivity losses to explore the effect of changing assumptions regarding the role of productivity losses. We will value productivity losses in the model by applying an average salary cost to estimated time off work as a result of intracranial bleeding.

We will undertake a literature review to estimate the effects of radiation exposure associated with radiological investigations (CT brain and skull radiography). We will then model these data to estimate a QALY loss and/or cost associated with each radiological investigation. This QALY loss and/or cost will then be applied to every patient in the model who receives a radiological investigation.

Analysis will be conducted in accordance with the NICE reference case.36 Net benefit analysis will be used to identify the most cost-effective option at varying thresholds of willingness to pay.37 The optimal strategy at the threshold currently used by NICE for decision-making will be presented as the optimal strategy for the NHS. The methodology used in the decision-analytic model will be dependent on the data that are available and the number of health states following the minor head injuries that are necessary to incorporate, with the most appropriate technique selected.

The exact modelling methodology to be used will be chosen once key data have been identified as attempting to manipulate data to fit a prespecified modelling structure will not be as accurate as choosing the method that can best represent the decision problem. The lead modeller has published papers using a wide range of decision methodologies, including discrete event simulation,38 meta-modelling,39 transition-state modelling,40 decision-tree modelling,35 and infectious disease modelling incorporating herd immunity,41 and we are confident that whatever modelling methodology is most appropriate will be able to be constructed. If possible, we shall attempt to calibrate the mathematical model with published data during the construction phase.

Probabilistic sensitivity analyses (PSAs) will be conducted in order that any interactions and non-linearities within the modelling are properly considered. Jack-knife techniques42 will be conducted to ensure that a sufficient number of PSA runs have been conducted to ensure that the average calculated from all runs for a management strategy is robust. Additionally the uncertainty associated in the actual mean net benefit will be provided using the percentile method in order that the full uncertainty in the results is reported. These analyses will facilitate the calculation of both full and partial expected value of perfect information, and if it is deemed appropriate an evaluation of the expected value of sample information will also be conducted.

The value of information analysis will help us to determine where funders of primary research in this important area (such as health technology assessment) should direct future studies to ensure that recommendations for policy and practice are more robust.

References

1.
National Institute for Health and Clinical Excellence (NICE). Head injury: triage, assessment, investigation and early management of head injury in infants, children and adults. London: NICE; 2003. NICE clinical guideline No. 4.
2.
National Institute for Health and Clinical Excellence (NICE). Head injury: triage, assessment, investigation and early management of head injury in infants, children and adults. London: NICE; 2007. NICE clinical guideline No. 56.
3.
Dunning J, Stratford-Smith P, Lecky F, Batchelor J, Hogg K, Browne J, et al. Emergency Medicine Research Group. A meta-analysis of clinical correlates that predict significant intracranial injury in adults with minor head trauma. J Neurotrauma. 2004;21:877–85. [PubMed: 15307900]
4.
Dunning J, Batchelor J, Stratford-Smith P, Teece S, Browne J, Sharpin C, et al. A meta-analysis of variables that predict significant intracranial injury in minor head trauma. Arch Dis Child. 2004;89:653–9. [PMC free article: PMC1719991] [PubMed: 15210499]
5.
Stiell IG, Wells GA, Vandemheen K, Clement C, Lesiuk H, Laupacis A, et al. The Canadian CT Head rule for patients with minor head injury. Lancet. 2001;357:1391–6. [PubMed: 11356436]
6.
Haydel MJ, Preston CA, Mills TJ, Luber S, Blaudeau E, DeBlieux PM. Indications for computed tomography in patients with minor head injury. N Engl J Med. 2000;343:100–5. [PubMed: 10891517]
7.
Richless LK, English K, Heller MB, Rachlin J, McClean P, Auble TE. A prospective evaluation of radiologic criteria for head injury patients in a community emergency department. Am J Emerg Med. 1993;11:327–30. [PubMed: 8216510]
8.
Miller EC, Holmes JF, Derlet RW. Utilizing clinical factors to reduce head CT scan ordering for minor head trauma patients. J Emerg Med. 1997;15:453–7. [PubMed: 9279694]
9.
Stiell IG, Clement CM, Rowe BH, Schull MJ, Brison R, Cass D, et al. Comparison of the Canadian CT Head rule and the New Orleans criteria in patients with minor head injury. JAMA. 2005;294:1511–18. [PubMed: 16189364]
10.
Smits M, Dippel DW, de Haan GG, Dekker HM, Vos PE, Kool DR, et al. External validation of the Canadian CT Head rule and New Orleans criteria for CT scanning in patients with minor head trauma. JAMA. 2005;294:1519–25. [PubMed: 16189365]
11.
Smits M, Dippel DW, Steyerberg EW, de Haan GG, Dekker HM, Vos PE, et al. Predicting intracranial traumatic findings on computed tomography in patients with minor head injury: the CHIP prediction rule. Ann Intern Med. 2007;146:397–405. [PubMed: 17371884]
12.
Oman JA, Cooper RJ, Holmes JF, Viccellio P, Nyce A, Ross SE, et al. Performance of a decision rule to predict need for CT among children with blunt head trauma. Pediatrics. 2006;117:e238–46. [PubMed: 16418311]
13.
Dunning J, Daly JP, Lomas JP, Lecky F, Batchelor K, Mackway-Jones K. Derivation of the children's head injury algorithm for the prediction of important clinical events decision rule for head injury in children. Arch Dis Child. 2006;91:885–91. [PMC free article: PMC2082967] [PubMed: 17056862]
14.
Hofman PA, Nelemans P, Kemerink GJ, Wilmink JT. Value of radiological diagnosis of skull fracture in the management of mild head injury: meta-analysis. J Neurol Neurosurg Psychiatry. 2000;68:416–22. [PMC free article: PMC1736859] [PubMed: 10727475]
15.
Kelly AB, Zimmerman RD, Snow RB, Gandy SE, Heier LBA, Deck MDF. Head trauma – comparison of MR and CT – experience of 100 patients. Am J Neuroradiol. 1988;9:699–708. [PubMed: 3135716]
16.
Ingebrigtsen T. Routine early CT-scan is cost saving after minor head injury. Acta Neurol Scand. 1996;93:207–10. [PubMed: 8741145]
17.
Af Geijerstam JL, Britton M, Marke LA. Minor head injury: observation or computer tomography? Economic aspects by literature review and economic analysis. Emerg Med J. 2004;21:54–8. [PMC free article: PMC1756360] [PubMed: 14734377]
18.
Norlund A, Marke LA, af Geijerstam JL, Oredsson S, Britton M., the OCTOPUS study investigators. Immediate computed tomography or admission for observation after mild head injury: cost comparison in randomised controlled trial. BMJ. 2006;333:469. [PMC free article: PMC1557926] [PubMed: 16895945]
19.
Af Geijerstam JL, Oredsson S, Britton M., the OCTOPUS study investigators. Medical outcome after immediate computed tomography or admission for observation in patients with mild head injury: randomised controlled trial. BMJ. 2006;333:465–8. [PMC free article: PMC1557917] [PubMed: 16895944]
20.
Stein SC. Indications for CT scanning in mild traumatic brain injury: a cost-effectiveness study. J Trauma. 2006;61:558–66. [PubMed: 16966987]
21.
Sultan HY, Boyle A, Pereira M, et al. Application of the Canadian CT head rules in managing minor head injuries in a UK emergency department: implications for the implementation of the NICE guidelines. Emerg Med J. 2004;21:420–5. [PMC free article: PMC1726397] [PubMed: 15208222]
22.
Hassan Z, Smith M, Littlewood S, Bouamra O, Hughes D, Biggin C, et al. Head injuries: a study evaluating the impact of the NICE head injury guidelines. Emerg Med J. 2005;22:845–9. [PMC free article: PMC1726640] [PubMed: 16299190]
23.
Shravat BP, Hynes KA. The impact of NICE guidelines for the management of head injury on the workload of the radiology department. Emerg Med J. 2004;21:521–2. [PMC free article: PMC1726374] [PubMed: 15208252]
24.
Goodacre S. Hospital admissions with head injury following publication of NICE guidance. Emerg Med J. 2008;25:556–7. [PubMed: 18723700]
25.
Westwood ME, Whiting PF, Kleijen J. How does study quality affect the results of a diagnostic meta-analysis? BMC Med Res Methodol. 2005;5:20. [PMC free article: PMC1180444] [PubMed: 15943861]
26.
Whiting P, Harbord R, Kleijnen J. No role for quality scores in systematic reviews of diagnostic accuracy studies. BMC Med Res Methodol. 2005;5:19. [PMC free article: PMC1184082] [PubMed: 15918898]
27.
Whiting P, Rutjes AW, Dinnes J, Reitsma JB, Bossuyt PM, Kleijnen J. A systematic review finds that diagnostic reviews fail to incorporate quality despite available tools. J Clin Epidemiol. 2005;58:1–12. [PubMed: 15649665]
28.
Whiting P, Rutjes AW, Dinnes J, Reitsma J, Bossuyt PM, Kleijnen J. Development and validation of methods for assessing the quality of diagnostic accuracy studies. Health Technol Assess. 2004;8(25) [PubMed: 15193208]
29.
Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: The STARD initiative. Department of Clinical Epidemiology and Biostatistics Academic Medical Centre, University of Amsterdam; the Netherlands: 2003.
30.
Lijmer JG, Mol JJ, Heisterkamp S, Bonsel GJ, van der Meulen JHP, Bossuyt PMM. Empirical evidence of design-related bias in studies of diagnostic tests. JAMA. 1999;282:1061–6. [PubMed: 10493205]
31.
Deeks JJ. Systematic reviews of evaluations of diagnostic and screening tests. In: Egger M, Davey Smith G, Altman DG, editors. Systematic reviews in health care: meta-analysis in context. 2nd edn. London: BMJ Publishing Group; 2001.
32.
Harbord RM, Deeks JJ, Egger M, Whiting P, Sterne JA. A unification of models for meta-analysis of diagnostic accuracy studies. Biostatistics. 2007;8:239–51. [PubMed: 16698768]
33.
Sutton AJ, Cooper NJ, Goodacre S, Stevenson M. Integration of meta-analysis and economic decision modelling for evaluating diagnostic tests. Med Decis Making. 2008;28:650–67. [PubMed: 18753686]
34.
Zamora J, Abraira V, Muriel A, Khan KS, Coomavasamy A. Meta-DiSc: a software for meta-analysis of test accuracy data. BMC Med Res Methodol. 2006;6:31. [PMC free article: PMC1552081] [PubMed: 16836745]
35.
Goodacre S, Sampson F, Stevenson M, Wailoo A, Sutton AJ, Thomas S, et al. Measurement of the clinical and cost-effectiveness of non-invasive diagnostic testing strategies for deep vein thrombosis. Health Technol Assess. 2006;10(15) [PubMed: 16707072]
36.
National Institute of Clinical Excellence (NICE). Guide to the methods of technology appraisal. London: NICE; 2004.
37.
Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost-effectiveness acceptability curves. Health Econ. 2001;10:779–87. [PubMed: 11747057]
38.
Stevenson MD, Oakley JE, Chick SE, Chalkidou K. The cost-effectiveness of surgical instrument management policies to reduce the risk of vCJD transmission to humans. JORS. 2009;60:506–18.
39.
Stevenson MD, Brazier JE, Calvert NW, Lloyd-Jones M, Oakley J, Kanis JA. Description of an individual patient methodology for calculating the cost-effectiveness of treatments for osteoporosis in women. JORS. 2005;56:214–21.
40.
Wardlaw JM, Chappell FM, Stevenson M, De Nigris E, Thomas S, Gillard J, et al. Accurate, practical and cost-effective assessment of carotid stenosis in the UK. Health Technol Assess. 2006;10(30) [PubMed: 16904049]
41.
Stevenson MD, Beard SM, Finn A, Brennan A. Estimating the potential health gain and cost consequences of introducing a pre-school pertussis booster into the UK child vaccination schedule. Vaccine. 2002;20:1778–86. [PubMed: 11906765]
42.
Inglehart DL. Simulating stable stochastic systems, V: comparison of ratio estimators. Naval Res Logist Quart. 1975;22:553–65.
© 2011, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Bookshelf ID: NBK99998

Views

Other titles in this collection

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...