Risk factors for readmission of inpatients with diabetes: A systematic review.

AIM: We have limited understanding of which risk factors contribute to increased readmission rates amongst people discharged from hospital with diabetes. We aim to complete the first review of its kind, to identify, in a systematic way, known risk factors for hospital readmission amongst people with diabetes, in order to better understand this costly complication. METHOD: The review was prospectively registered in the PROSPERO database. Risk factors were identified through systematic review of literature in PubMed, EMBASE & SCOPUS databases, performed independently by two authors prior to data extraction, with quality assessment and semi-quantitative synthesis according to PRISMA guidelines. RESULTS: Eighty-three studies were selected for inclusion, predominantly from the United States, and utilising retrospective analysis of local or regional data sets. 76 distinct statistically significant risk factors were identified across 48 studies. The most commonly identified risk factors were; co-morbidity burden, age, race and insurance type. Few studies conducted power calculations; unstandardized effect sizes were calculated for the majority of statistically significant risk factors. CONCLUSION: This review is important in assessing the current state of the literature and in supporting development of interventions to reduce readmission risk. Furthermore, it provides an important foundation for development of rigorous, pre-specified risk prediction models.

Clinical data integration model. Core interoperability ontology for research using primary care data.

INTRODUCTION: This article is part of the Focus Theme of METHODS of Information in Medicine on "Managing Interoperability and Complexity in Health Systems". BACKGROUND: Primary care data is the single richest source of routine health care data. However its use, both in research and clinical work, often requires data from multiple clinical sites, clinical trials databases and registries. Data integration and interoperability are therefore of utmost importance. OBJECTIVES: TRANSFoRm's general approach relies on a unified interoperability framework, described in a previous paper. We developed a core ontology for an interoperability framework based on data mediation. This article presents how such an ontology, the Clinical Data Integration Model (CDIM), can be designed to support, in conjunction with appropriate terminologies, biomedical data federation within TRANSFoRm, an EU FP7 project that aims to develop the digital infrastructure for a learning healthcare system in European Primary Care. METHODS: TRANSFoRm utilizes a unified structural / terminological interoperability framework, based on the local-as-view mediation paradigm. Such an approach mandates the global information model to describe the domain of interest independently of the data sources to be explored. Following a requirement analysis process, no ontology focusing on primary care research was identified and, thus we designed a realist ontology based on Basic Formal Ontology to support our framework in collaboration with various terminologies used in primary care. RESULTS: The resulting ontology has 549 classes and 82 object properties and is used to support data integration for TRANSFoRm's use cases. Concepts identified by researchers were successfully expressed in queries using CDIM and pertinent terminologies. As an example, we illustrate how, in TRANSFoRm, the Query Formulation Workbench can capture eligibility criteria in a computable representation, which is based on CDIM. CONCLUSION: A unified mediation approach to semantic interoperability provides a flexible and extensible framework for all types of interaction between health record systems and research systems. CDIM, as core ontology of such an approach, enables simplicity and consistency of design across the heterogeneous software landscape and can support the specific needs of EHR-driven phenotyping research using primary care data.

(1)H magnetic resonance spectroscopy in the diagnosis of paediatric low grade brain tumours.

INTRODUCTION: Low grade gliomas are the commonest brain tumours in children but present in a myriad of ways, each with its own treatment challenges. Conventional MRI scans play an important role in their management but have limited ability to identify likely clinical behaviour. The aim of this study is to investigate (1)H magnetic resonance spectroscopy (MRS) as a method for detecting differences between the various low grade gliomas and related tumours in children. PATIENTS AND METHODS: Short echo time single voxel (1)H MRS at 1.5 or 3.0 T was performed prior to treatment on children with low grade brain tumours at two centres and five MR scanners, 69 cases had data which passed quality control. MRS data was processed using LCModel to give mean spectra and metabolite concentrations which were compared using T-tests, ANOVA, Receiver Operator Characteristic curves and logistic regression in SPSS. RESULTS: Significant differences were found in concentrations of key metabolites between glioneuronal and glial tumours (T-test p<0.05) and between most of the individual histological subtypes of low grade gliomas. The discriminatory metabolites identified, such as choline and myoinositol, are known tumour biomarkers. In the set of pilocytic astrocytomas and unbiopsied optic pathway gliomas, significant differences (p<0.05, ANOVA) were found in metabolite profiles of tumours depending on location and patient neurofibromatosis type 1 status. Logistic regression analyses yielded equations which could be used to assess the probability of a tumour being of a specific type. CONCLUSIONS: MRS can detect subtle differences between low grade brain tumours in children and should form part of the clinical assessment of these tumours.

A systematic review of positron emission tomography (PET) and positron emission tomography/computed tomography (PET/CT) for the diagnosis of breast cancer recurrence.

BACKGROUND: Breast cancer (BC) accounts for one-third of all cases of cancer in women in the UK. Current strategies for the detection of BC recurrence include computed tomography (CT), magnetic resonance imaging (MRI) and bone scintigraphy. Positron emission tomography (PET) and, more recently, positron emission tomography/computed tomography (PET/CT) are technologies that have been shown to have increasing relevance in the detection and management of BC recurrence. OBJECTIVE: To review the accuracy of PET and PET/CT for the diagnosis of BC recurrence by assessing their value compared with current practice and compared with each other. DATA SOURCES: MEDLINE and EMBASE were searched from inception to May 2009. STUDY SELECTION: Studies were included if investigations used PET or PET/CT to diagnose BC recurrence in patients with a history of BC and if the reference standard used to define the true disease status was histological diagnosis and/or long-term clinical follow-up. Studies were excluded if a non-standard PET or PET/CT technology was used, investigations were conducted for screening or staging of primary breast cancer, there was an inadequate or undefined reference standard, or raw data for calculation of diagnostic accuracy were not available. STUDY APPRAISAL: Quality assessment and data extraction were performed independently by two reviewers. Direct and indirect comparisons were made between PET and PET/CT and between these technologies and methods of conventional imaging, and meta-analyses were carried out. Analysis was conducted separately on patient- and lesion-based data. Subgroup analysis was conducted to investigate variation in the accuracy of PET in certain populations or contexts and sensitivity analysis was conducted to examine the reliability of the primary outcome measures. RESULTS: Of the 28 studies included in the review, 25 presented patient-based data and 7 presented lesion-based data for PET and 5 presented patient-based data and 1 presented patient- and lesion-based data for PET/CT; 16 studies conducted direct comparisons with 12 comparing the accuracy of PET or PET/CT with conventional diagnostic tests and 4 with MRI. For patient-based data (direct comparison) PET had significantly higher sensitivity [89%, 95% confidence interval (CI) 83% to 93% vs 79%, 95% CI 72% to 85%, relative sensitivity 1.12, 95% CI 1.04 to 1.21, p = 0.005] and significantly higher specificity (93%, 95% CI 83% to 97% vs 83%, 95% CI 67% to 92%, relative specificity 1.12, 95% CI 1.01 to 1.24, p = 0.036) compared with conventional imaging tests (CITs)--test performance did not appear to vary according to the type of CIT tested. For patient-based data (direct comparison) PET/CT had significantly higher sensitivity compared with CT (95%, 95% CI 88% to 98% vs 80%, 95% CI 65% to 90%, relative sensitivity 1.19, 95% CI 1.03 to 1.37, p = 0.015), but the increase in specificity was not significant (89%, 95% CI 69% to 97% vs 77%, 95% CI 50% to 92%, relative specificity 1.15, 95% CI 0.95 to 1.41, p = 0.157). For patient-based data (direct comparison) PET/CT had significantly higher sensitivity compared with PET (96%, 95% CI 90% to 98% vs 85%, 95% CI 77% to 91%, relative sensitivity 1.11, 95% CI 1.03 to 1.18, p = 0.006), but the increase in specificity was not significant (89%, 95% CI 74% to 96% vs 82%, 95% CI 64% to 92%, relative specificity 1.08, 95% CI 0.94 to 1.20, p = 0.267). For patient-based data there were no significant differences in the sensitivity or specificity of PET when compared with MRI, and, in the one lesion based study, there was no significant differences in the sensitivity or specificity of PET/CT when compared with MRI. LIMITATIONS: Studies reviewed were generally small and retrospective and this may have limited the generalisability of findings. Subgroup analysis was conducted on the whole set of studies investigating PET and was not restricted to comparative studies. Conventional imaging studies that were not compared with PET or PET/CT were excluded from the review. CONCLUSIONS: Available evidence suggests that for the detection of BC recurrence PET, in addition to conventional imaging techniques, may generally offer improved diagnostic accuracy compared with current standard practice. However, uncertainty remains around its use as a replacement for, rather than an add-on to, existing imaging technologies. In addition, PET/CT appeared to show clear advantage over CT and PET alone for the diagnosis of BC recurrence. FUTURE WORK: Future research should include: prospective studies with patient populations clearly defined with regard to their clinical presentation; a study of diagnostic accuracy of PET/CT compared with conventional imaging techniques; a study of PET/CT compared with whole-body MRI; studies investigating the possibility of using PET/CT as a replacement for rather than an addition to CITs; and using modelling of the impact of PET/CT on patient outcomes to inform the possibility of conducting large-scale intervention trials.

Non-invasive detection of glycine as a biomarker of malignancy in childhood brain tumours using in-vivo 1H MRS at 1.5 tesla confirmed by ex-vivo high-resolution magic-angle spinning NMR.

Management of brain tumours in children would benefit from improved non-invasive diagnosis, characterisation and prognostic biomarkers. Metabolite profiles derived from in-vivo MRS have been shown to provide such information. Studies indicate that using optimum a priori information on metabolite contents in the construction of linear combination (LC) models of MR spectra leads to improved metabolite profile estimation. Glycine (Gly) is usually neglected in such models due to strong overlap with myo-inositol (mI) and a low concentration in normal brain. However, biological studies indicate that Gly is abundant in high-grade brain tumours. This study aimed to investigate the quantitation of Gly in paediatric brain tumours using MRS analysed by LCModel, and its potential as a non-invasive biomarker of malignancy. Single-voxel MRS was performed using PRESS (TR 1500 ms, TE 30 ms/135 ms) on a 1.5 T scanner. Forty-seven cases (18 high grade (HG), 17 low grade (LG), 12 ungraded) were retrospectively selected if both short-TE and long-TE MRS (n = 33) or short-TE MRS and high-resolution magic-angle spinning (HRMAS) of matched surgical samples (n = 15) were available. The inclusion of Gly in LCModel analyses led to significantly reduced fit residues for both short-TE and long-TE MRS (p < 0.05). The Gly concentrations estimated from short-TE MRS were significantly correlated with the long-TE values (R = 0.91, p < 0.001). The Gly concentration estimated by LCModel was significantly higher in HG versus LG tumours for both short-TE (p < 1e-6) and long-TE (p = 0.003) MRS. This was consistent with the HRMAS results, which showed a significantly higher normalised Gly concentration in HG tumours (p < 0.05) and a significant correlation with the normalised Gly concentration measured from short-TE in-vivo MRS (p < 0.05). This study suggests that glycine can be reliably detected in paediatric brain tumours using in-vivo MRS on standard clinical scanners and that it is a promising biomarker of tumour aggressiveness.

Identification and characterisation of childhood cerebellar tumours by in vivo proton MRS.

(1)H MRS has great potential for the clinical investigation of childhood brain tumours, but the low incidence in, and difficulties of performing trials on, children have hampered progress in this area. Most studies have used a long-TE, thus limiting the metabolite information obtained, and multivariate analysis has been largely unexplored. Thirty-five children with untreated cerebellar tumours (18 medulloblastomas, 12 pilocytic astrocytomas and five ependymomas) were investigated using a single-voxel short-TE PRESS sequence on a 1.5 T scanner. Spectra were analysed using LCModel to yield metabolite profiles, and key metabolite assignments were verified by comparison with high-resolution magic-angle-spinning NMR of representative tumour biopsy samples. In addition to univariate metabolite comparisons, the use of multivariate classifiers was investigated. Principal component analysis was used for dimension reduction, and linear discriminant analysis was used for variable selection and classification. A bootstrap cross-validation method suitable for estimating the true performance of classifiers in small datasets was used. The discriminant function coefficients were stable and showed that medulloblastomas were characterised by high taurine, phosphocholine and glutamate and low glutamine, astrocytomas were distinguished by low creatine and high N-acetylaspartate, and ependymomas were differentiated by high myo-inositol and glycerophosphocholine. The same metabolite features were seen in NMR spectra of ex vivo samples. Successful classification was achieved for glial-cell (astrocytoma + ependymoma) versus non-glial-cell (medulloblastoma) tumours, with a bootstrap 0.632 + error, e(B.632+), of 5.3%. For astrocytoma vs medulloblastoma and astrocytoma vs medulloblastoma vs ependymoma classification, the e(B.632+) was 6.9% and 7.1%, respectively. The study showed that (1)H MRS detects key differences in the metabolite profiles for the main types of childhood cerebellar tumours and that discriminant analysis of metabolite profiles is a promising tool for classification. The findings warrant confirmation by larger multi-centre studies.

Structural neuroimaging in psychosis: a systematic review and economic evaluation.

OBJECTIVES: To establish the clinical effectiveness and cost-effectiveness of structural neuroimaging [structural magnetic resonance imaging (MRI) or computed tomography (CT) scanning] for all patients with psychosis, particularly a first episode of psychosis, relative to the current UK practice of selective screening only where it is clinically indicated. DATA SOURCES: Major electronic databases were searched from inception to November 2006. REVIEW METHODS: A systematic review of studies reporting the additional diagnostic benefit of structural MRI, CT or combinations of these in patients with psychosis was conducted. The economic assessment consisted of a systematic review of economic evaluations and the development of a threshold analysis to predict the gain in quality-adjusted life-years (QALYs) required to make neuroimaging cost-effective at commonly accepted threshold levels (20,000 pounds and 30,000 pounds per QALY). Sensitivity analyses of several parameters including prevalence of psychosis were performed. RESULTS: The systematic review included 24 studies of a diagnostic before-after type of design evaluating the clinical benefit of CT, structural MRI or combinations in treatment-naive, first-episode or unspecified psychotic patients, including one in schizophrenia patients resistant to treatment. Also included was a review of published case reports of misidentification syndromes. Almost all evidence was in patients aged less than 65 years. In most studies, structural neuroimaging identified very little that would influence patient management that was not suspected based on a medical history and/or physical examination and there were more incidental findings. In the four MRI studies, approximately 5% of patients had findings that would influence clinical management, whereas in the CT studies, approximately 0.5% of patients had these findings. The review of misidentification syndromes found that 25% of CT scans affected clinical management, but this may have been a selected and therefore unrepresentative sample. A threshold analysis with a 1-year time horizon was undertaken. This combined the incremental cost of routine scanning with a threshold cost per QALY value of 20,000 pounds and 30,000 pounds to predict the QoL gain required to meet these threshold values. Routine scanning versus selective scanning appears to produce different results for MRI and CT. With MRI scanning the incremental cost is positive, ranging from 37 pounds to 150 pounds; however, when scanning routinely using CT, the result is cost saving, ranging from 7 pounds to 108 pounds with the assumption of a 1% prevalence rate of tumours/cysts or other organic causes amenable to treatment. This means that for the intervention to be viewed as cost-effective, the QALY gain necessary for MRI scanning is 0.002-0.007 and with CT scanning the QALY loss that can be tolerated is between 0.0003 and 0.0054 using a 20,000 pounds threshold value. These estimates were subjected to sensitivity analysis. With a 3-month time delay, MRI remains cost-incurring with a small gain in QoL required for the intervention to be cost-effective; routine scanning with CT remains cost-saving. When the sensitivity of CT is varied to 50%, routine scanning is both cost-incurring or cost-saving depending on the scenario. Finally, the results have been shown to be sensitive to the assumed prevalence rate of brain tumours in a psychotic population. CONCLUSIONS: The evidence to date suggests that if screening with structural neuroimaging was implemented in all patients presenting with psychotic symptoms, little would be found to affect clinical management in addition to that suspected by a full clinical history and neurological examination. From an economic perspective, the outcome is not clear. The strategy of neuroimaging for all is either cost-incurring or cost-saving (dependent upon whether MRI or CT is used) if the prevalence of organic causes is around 1%. However, these values are nested within a number of assumptions, and so have to be interpreted with caution. The main research priorities are to monitor the current use of structural neuroimaging in psychosis in the NHS to identify clinical triggers to its current use and subsequent outcomes; to undertake well-conducted diagnostic before-and-after studies on representative populations to determine the clinical utility of structural neuroimaging in this patient group, and to determine whether the most appropriate structural imaging modality in psychosis should be CT or MRI.

A multistage perceptual quality assessment for compressed digital angiogram images.

This paper describes a multistage perceptual quality assessment (MPQA) model for compressed images. The motivation for the development of a perceptual quality assessment is to measure (in)visible differences between original and processed images. The MPQA produces visible distortion maps and quantitative error measures informed by considerations of the human visual system (HVS). Original and decompressed images are decomposed into different spatial frequency bands and orientations modeling the human cortex. Contrast errors are calculated for each frequency and orientation, and masked as a function of contrast sensitivity and background uncertainty. Spatially masked contrast error measurements are then made across frequency bands and orientations to produce a single perceptual distortion visibility map (PDVM). A perceptual quality rating (PQR) is calculated from the PDVM and transformed into a one to five scale, PQR(1-5), for direct comparison with the mean opinion score, generally used in subjective ratings. The proposed MPQA model is based on existing perceptual quality assessment models, while it is differentiated by the inclusion of contrast masking as a function of background uncertainty. A pilot study of clinical experiments on wavelet-compressed digital angiogram has been performed on a sample set of angiogram images to identify diagnostically acceptable reconstruction. Our results show that the PQR(1-5) of diagnostically acceptable lossy image reconstructions have better agreement with cardiologists' responses than objective error measurement methods, such as peak signal-to-noise ratio A Perceptual thresholding and CSF-based Uniform quantization (PCU) method is also proposed using the vision models presented in this paper. The vision models are implemented in the thresholding and quantization stages of a compression algorithm and shown to produce improved compression ratio performance with less visible distortion than that of the embedded zerotrees wavelet (EZWs).

WWW creates new interactive 3D graphics and collaborative environments for medical research and education.

Virtual Reality Modelling Language (VRML) is the start of a new era for medicine and the World Wide Web (WWW). Scientists can use VRML across the Internet to explore new three-dimensional (3D) worlds, share concepts and collaborate together in a virtual environment. VRML enables the generation of virtual environments through the use of geometric, spatial and colour data structures to represent 3D objects and scenes. In medicine, researchers often want to interact with scientific data, which in several instances may also be dynamic (e.g. MRI data). This data is often very large and is difficult to visualise. A 3D graphical representation can make the information contained in such large data sets more understandable and easier to interpret. Fast networks and satellites can reliably transfer large data sets from computer to computer. This has led to the adoption of remote tale-working in many applications including medical applications. Radiology experts, for example, can view and inspect in near real-time a 3D data set acquired from a patient who is in another part of the world. Such technology is destined to improve the quality of life for many people. This paper introduces VRML (including some technical details) and discusses the advantages of VRML in application developing.

An intranet database for pacemaker patients.

A database system, incorporating smartcard technologies, was designed to hold the personal and pacing details of pacemaker patients, who attended a clinic at the Royal Sussex County Hospital (RSCH), Brighton, UK. Following an initial period of a 12 month clinical trial, with the database running on a standalone personal computer, the Pacemaker Patient Database has been redesigned and implemented as an intranet-based system. This paper describes the issues relating to the development of the new prototype system and identifies the design principles for intranet-based electronic health care (EHCR) record database systems.

Automated feature extraction for the classification of human in vivo 13C NMR spectra using statistical pattern recognition and wavelets.

If magnetic resonance spectroscopy (MRS) is to become a useful tool in clinical medicine, it will be necessary to find reliable methods for analyzing and classifying MRS data. Automated methods are desirable because they can remove user bias and can deal with large amounts of data, allowing the use of all the available information. In this study, techniques for automatically extracting features for the classification of MRS in vivo data are investigated. Among the techniques used were wavelets, principal component analysis, and linear discriminant function analysis. These techniques were tested on a set of 75 in vivo 13C spectra of human adipose tissue from subjects from three different dietary groups (vegan, vegetarian, and omnivore). It was found that it was possible to assign automatically 94% of the vegans and omnivores to their correct dietary groups, without the need for explicit identification or measurement of peaks.

Reject analysis: a pilot programme for image quality management.

The radiographic film wastage and the different parameters affecting this wastage were analysed for a 9-week period at a 600-bed University Hospital. An overall reject rate of 7.6% was found. The different reasons for rejection were evaluated, while retake rate, relation between working experience of the personnel, amount of rejected films and total film wastage in surface (m2), were registered and analysed.