Extreme regression models for characterizing high-cost patients

OBJECTIVE

Healthcare cost distribution generally presents a high level of skewness, with a relatively small number of subjects accounting for a large portion of healthcare expenditures. Information on factors that predict high expenditures is of interest in healthcare planning. The aim of this paper was to inspect the behaviour of extreme regression (ER) models.

METHODS

We performed a simple simulation study, based on the LogNormal distribution, to assess the performance of ER in the special cases of heterogeneity and strong asymmetry of the cost variable. We then discussed the application of ER models to the analysis of three data sets of diabetes, lung cancer and myocardial infarction patients.

RESULTS

The ER showed to be able to cope fairly well with skewed distribution but under the condition that all observations have strictly positive costs.

CONCLUSION

The main advantage of the ER model is to unify these approaches in a unique framework, where the estimation of the cut-offs and the production of the prediction rules are performed simultaneously for a continuous response variable. The final model can thus be analysed at any desiderate quantile of the cost distribution, avoiding the need of pre-specifying any cut-off.

Relative survival in dilated cardiomyopathy: a stratification study of long-term outcome to evaluate life insurance cover

OBJECTIVES

The aim of this study was to compare mortality of dilated cardiomyopathy (DCM) patients with the mortality in the background Italian population, taking into account demographic characteristics and clinical stratification of long-term outcome, ie, "reverse remodelling" within the first 2 years of follow-up.

BACKGROUND

DCM is a myocardial disease, characterized by left and/or right ventricular dilation and dysfunction and poor outcome. Evidence-based treatment with ACE inhibitors, beta-blockers and, in the last decade, implantable cardioverter defibrillators have been demonstrated to improve significantly heart failure symptoms and prognosis. At present, DCM patients are unlikely to be accepted for life insurance.

METHODS

A cohort of 577 DCM patients consecutively enrolled from 1988 to 2004 in the Heart Muscle Disease Registry of Trieste, Italy, was matched by sex, age and registry data entry with the mortality data of the Italian population. Relative survival has been estimated by means of Kaplan-Meier technique, and mortality ratios (MR) with corresponding 95% confidence intervals have been computed.

RESULTS

DCM patients who showed a significant reverse remodelling within the first 2 years of treatment showed comparable survival with respect to the control population, and therefore could be taken into consideration for life insurance coverage, at least for a short or medium-term of years.

CONCLUSIONS

The data illustrate that survival probability strongly depend on the individual treatment and evolution of the disease and could be easily measured within the first 2 years of follow-up. If this information is collected at the time of evaluation of an applicant for life cover, the insurance company could possibly improve its risk stratification.

Could birthweight prediction models be improved by adding fetal subcutaneous tissue thickness?

AIM

The aims of the study were to: (i) compare the accuracy of standard ultrasonic algorithms in the estimation of fetal weight and; (ii) test two new algorithms in order to improve the global performance of birthweight prediction by adding fetal subcutaneous tissue thickness.

METHODS

We enrolled 398 patients who were between 34 and 42 weeks' gestation. Routine ultrasonographic biometric parameters as well as subcutaneous tissue thickness ultrasound parameters were measured. Correlation matrices between ultrasound parameters, in order to evaluate the degree of multicollinearity between these parameters, were assessed to develop a stepwise multiple regression birthweight predictive model.

RESULTS

Contributions of single ultrasound measurements in predicting birthweight were examined, by fitting Log-transformed birthweight versus single ultrasound measurements. We found that the mid-thigh tissue area was able to significantly improve the performance of the birthweight prediction process when added to the other standard ultrasound measurements. We derived two new algorithms which appeared to be better at predicting birthweight. Furthermore there was a lower minimum absolute estimation error noted when compared to other reported formulae.

CONCLUSIONS

Our algorithms showed that the addition of the mid-thigh tissue evaluation in birthweight prediction was valuable in comparison to birthweight prediction models which are based on routine ultrasound parameters.

Hand sensory-motor cortical network assessed by functional source separation

The functional source separation procedure (FSS) was applied to identify the activities of the primary sensorimotor areas (SM1) devoted to hand control. FSS adds a functional constraint to the cost function of the basic independent component analysis, and obtains source activity all along different processing states. Magnetoencephalographic signals from the left SM1 were recorded in 14 healthy subjects during a simple sensorimotor paradigm--galvanic right median nerve stimuli intermingled with submaximal isometric thumb opposition. Two functional sources related to the sensory flow in the primary cortex were extracted requiring maximal responsiveness to the nerve stimulation at around 20 and 30 ms (S1a, S1b). Maximal cortico-muscular coherence was required for the extraction of the motor source (M1). Sources were multiplied by the Euclidean norm of their corresponding weight vectors, allowing amplitude comparisons among sources in a fixed position. In all subjects, S1a, S1b, M1 were successfully obtained, positioned consistently with the SM1 organization, and behaved as physiologically expected during the movement and processing of the sensory stimuli. The M1 source reacted to the nerve stimulation with higher intensity at latencies around 30 ms than around 20 ms. The FSS method was demonstrated to be able to obtain the dynamics of different primary cortical network activities, two devoted mainly to sensory inflow, and the other to the motor control of the contralateral hand. It was possible to observe each source both during pure sensory processing and during motor tasks. In all conditions, a direct comparison of source intensities can be achieved.

Functional source separation applied to induced visual gamma activity

Objective of this work was to explore the performance of a recently introduced source extraction method, FSS (Functional Source Separation), in recovering induced oscillatory change responses from extra-cephalic magnetoencephalographic (MEG) signals. Unlike algorithms used to solve the inverse problem, FSS does not make any assumption about the underlying biophysical source model; instead, it makes use of task-related features (functional constraints) to estimate source/s of interest. FSS was compared with blind source separation (BSS) approaches such as Principal and Independent Component Analysis, PCA and ICA, which are not subject to any explicit forward solution or functional constraint, but require source uncorrelatedness (PCA), or independence (ICA). A visual MEG experiment with signals recorded from six subjects viewing a set of static horizontal black/white square-wave grating patterns at different spatial frequencies was analyzed. The beamforming technique Synthetic Aperture Magnetometry (SAM) was applied to localize task-related sources; obtained spatial filters were used to automatically select BSS and FSS components in the spatial area of interest. Source spectral properties were investigated by using Morlet-wavelet time-frequency representations and significant task-induced changes were evaluated by means of a resampling technique; the resulting spectral behaviours in the gamma frequency band of interest (20-70 Hz), as well as the spatial frequency-dependent gamma reactivity, were quantified and compared among methods. Among the tested approaches, only FSS was able to estimate the expected sustained gamma activity enhancement in primary visual cortex, throughout the whole duration of the stimulus presentation for all subjects, and to obtain sources comparable to invasively recorded data.

Ultrasound-determined fetal subcutaneous tissue thickness for a birthweight prediction model

AIM

The aim of this study was to explore a birthweight prediction model using ultrasound determined tissue thickness (SCTT) parameters.

METHODS

We measured routine ultrasonographic biometric parameters and in addition, fetal SCTT in 201 healthy singleton pregnancies. Mid-arm fat and lean mass, mid-thigh fat and lean mass, subscapular fat mass and abdominal fat mass (AFM) were measured in order to calculate a birthweight prediction model. Ultrasound measurements were analyzed using an 'anovarepeated measures model'. The growth rate (beta-slope) of the selected parameters was computed and the correlation coefficient with the birthweight and the Kendall rank correlation tau, were calculated.

RESULTS

From the ultrasound determined SCTT parameters, only abdominal circumference (AC), AFM, and MTLM showed a statistically significant trend. The beta-slope of mid-thigh lean mass was excluded since it exhibited significant correlation with the beta-slope of AFM. The final regression model could be calculated as: birthweight (gr.) = intercept +alpha(1)(AFM beta-slope) + alpha(2)(AC beta-slope), where alpha(1), alpha(2) represent regression coefficients.

CONCLUSIONS

We provide a graphical birthweight prediction model for clinical practice using conventional and specific ultrasound measurements of fetal subcutaneous tissue thickness. This model is based upon an overall analysis of the ultrasound estimated body components.

Single inherited thrombophilias and adverse pregnancy outcomes

INTRODUCTION

Inherited thrombophilia is believed to be a multiple gene disease with more than one defect. We aimed to determine the association between single thrombophilic patterns and a variety of pregnancy diseases.

METHODS

284 pregnant women were recruited for the present study and were divided in two groups: A group (176 controls) and B group (108 cases). Patients belonging to the B group had one of the following: severe pre-eclampsia, hemolysis, hepatic enzymes increase, hypertension and low platelet count (HELLP) syndrome, gestational hypertension, fetal growth restriction, intrauterine death, abruptio placentae and disseminated intravascular coagulopathy. To detect methylenetetrahydrofolate reductase (MTHFR) A1298C, MTHFR C677T, factor V Leiden, PAI-1, mutant prothrombin G20210A, an inverse hybridization technology was used. Plasma homocysteine, antithrombin (AT) III and protein S were determined. A modified functional activated protein C resistance was detected.

RESULTS

MTHFR C677T and hyperhomocysteinemia were more prevalent than other thrombophilias. Deficiency in AT III was significantly linked with pre-eclampsia (relative risk 0.88; 95% CI 0.83-0.94). Activated protein C resistance (APCR) was significantly related to the abruptio placentae (relative risk 0.71; 95% CI 0.61-0.82).

COMMENTS

Apart from the linkage between AT III deficiency and the occurrence of pre-eclampsia, and apart from the increased risk of abruptio placentae in pregnant women with altered APCR, we obtained findings in contrast with some of the published literature. In our case series, no association of pre-eclampsia with factor V Leiden or with prothrombin gene mutation was found.

‘Free’ copper in serum of Alzheimer’s disease patients correlates with markers of liver function

Non-ceruloplasmin bound copper ('free') seems slightly elevated in Alzheimer's disease (AD) patients. To test the hypothesis of a correlation between 'free' copper and liver function in AD. We evaluated 51 AD patients and 53 controls through typical tests for chronic liver disease (AST, ALT, gamma-GT, Albumin, prothrombin time - PT-, bilirubins), along with copper, ceruloplasmin, iron, cholesterol in the serum and apolipoprotein E epsilon4 (APOE4) genotype. Absolute serum copper and 'free' copper were higher, albumin was lower and PT longer in AD patients than in controls. 'Free' copper correlated negatively with markers of liver function, in that albumin and albumin/PT ratio (r = -0.43, p = 0.004), and positively with direct bilirubin. Copper and 'free' copper were higher in the APOE4 carriers. These results suggest that abnormalities in copper metabolism might have an effect on liver function in AD.

Ictal clinical and scalp-EEG findings differentiating temporal lobe epilepsies from temporal 'plus' epilepsies

Temporal 'plus' epilepsies are characterized by seizures involving a complex epileptogenic network including the temporal lobe and the closed neighboured structures such as the orbito-frontal cortex, the insula, the frontal and parietal operculum and the temporo-parieto-occipital junction. Temporal 'plus' epilepsies are currently identified by means of intracerebral electrodes but whether their diagnosis can be suspected non-invasively has not been evaluated yet. The aim of this retrospective study was to address this issue in 80 consecutive patients who were thought to suffer from non-lesional temporal lobe seizures which finally proved, on the basis of stereotactic intracerebral EEG (SEEG) recordings, to be 'purely' temporal (TL group, n = 58) or temporal 'plus' (T+ group, n = 22). Our results showed that the two groups of patients were difficult to differentiate on the basis of general clinical features or MRI data. Even the presence of hippocampal sclerosis did not distinguish the two groups. Conversely, both ictal clinical symptoms and scalp-EEG findings significantly differentiated TL from T+ patients. Patients with TL epilepsies more frequently presented an ability to warn at seizure onset (P = 0.003), an abdominal aura (P = 0.05), gestural automatisms (P = 0.04) and a post-ictal amnesia (P = 0.02). Patients suffering from T+ epilepsies more frequently had gustatory hallucinations (P = 0.02), rotatory vertigo (P = 0.02) and auditory illusions (P = 0.02) at seizure onset; they exhibited more frequently contraversive manifestations of the eyes and/or head (P = 0.001), piloerection (P = 0.03) and ipsilateral tonic motor signs (P = 0.05), and they were more often dysphoric in the post-ictal phase (P = 0.0001). Cluster analysis mainly indicated that some associations of symptoms were relevant for differentiating TL cases from T+ cases. Interictal EEG of T+ patients more frequently exhibited bilateral or precentral abnormalities, while ictal EEG more frequently pointed over the anterior frontal, temporo-parietal and precentral regions. Neither TL interictal spikes, nor TL ictal EEG onset, allowed us definitely to rule out the possibility of T+ epilepsies. Our findings may be useful for identifying, among patients suffering from 'atypical' non-lesional TL epilepsies, those who should undergo invasive recordings before surgery.

Functional source separation and hand cortical representation for a brain-computer interface feature extraction

A brain-computer interface (BCI) can be defined as any system that can track the person's intent which is embedded in his/her brain activity and, from it alone, translate the intention into commands of a computer. Among the brain signal monitoring systems best suited for this challenging task, electroencephalography (EEG) and magnetoencephalography (MEG) are the most realistic, since both are non-invasive, EEG is portable and MEG could provide more specific information that could be later exploited also through EEG signals. The first two BCI steps require set up of the appropriate experimental protocol while recording the brain signal and then to extract interesting features from the recorded cerebral activity. To provide information useful in these BCI stages, our aim is to provide an overview of a new procedure we recently developed, named functional source separation (FSS). As it comes from the blind source separation algorithms, it exploits the most valuable information provided by the electrophysiological techniques, i.e. the waveform signal properties, remaining blind to the biophysical nature of the signal sources. FSS returns the single trial source activity, estimates the time course of a neuronal pool along different experimental states on the basis of a specific functional requirement in a specific time period, and uses the simulated annealing as the optimization procedure allowing the exploit of functional constraints non-differentiable. Moreover, a minor section is included, devoted to information acquired by MEG in stroke patients, to guide BCI applications aiming at sustaining motor behaviour in these patients. Relevant BCI features - spatial and time-frequency properties - are in fact altered by a stroke in the regions devoted to hand control. Moreover, a method to investigate the relationship between sensory and motor hand cortical network activities is described, providing information useful to develop BCI feedback control systems. This review provides a description of the FSS technique, a promising tool for the BCI community for online electrophysiological feature extraction, and offers interesting information to develop BCI applications to sustain hand control in stroke patients.

Functional source separation from magnetoencephalographic signals

We propose a novel cerebral source extraction method (functional source separation, FSS) starting from extra-cephalic magnetoencephalographic (MEG) signals in humans. It is obtained by adding a functional constraint to the cost function of a basic independent component analysis (ICA) model, defined according to the specific experiment under study, and removing the orthogonality constraint, (i.e., in a single-unit approach, skipping decorrelation of each new component from the subspace generated by the components already found). Source activity was obtained all along processing of a simple separate sensory stimulation of thumb, little finger, and median nerve. Being the sources obtained one by one in each stage applying different criteria, the a posteriori "interesting sources selection" step is avoided. The obtained solutions were in agreement with the homuncular organization in all subjects, neurophysiologically reacting properly and with negligible residual activity. On this basis, the separated sources were interpreted as satisfactorily describing highly superimposed and interconnected neural networks devoted to cortical finger representation. The proposed procedure significantly improves the quality of the extraction with respect to a standard BSS algorithm. Moreover, it is very flexible in including different functional constraints, providing a promising tool to identify neuronal networks in very general cerebral processing.

Recognition of schematic facial displays of emotion in parents of children with autism

Performance on an emotional labeling task in response to schematic facial patterns representing five basic emotions without the concurrent presentation of a verbal category was investigated in 40 parents of children with autism and 40 matched controls. 'Autism fathers' performed worse than 'autism mothers', who performed worse than controls in decoding displays representing sadness or disgust. This indicates the need to include facial expression decoding tasks in genetic research of autism. In addition, emotional expression interactions between parents and their children with autism, particularly through play, where affect and prosody are 'physiologically' exaggerated, may stimulate development of social competence. Future studies could benefit from a combination of stimuli including photographs and schematic drawings, with and without associated verbal categories. This may allow the subdivision of patients and relatives on the basis of the amount of information needed to understand and process social-emotionally relevant information.

Excess of nonceruloplasmin serum copper in AD correlates with MMSE, CSF [beta]-amyloid, and h-tau

OBJECTIVE

To assess whether serum copper in Alzheimer disease (AD) correlates with cognitive scores, beta-amyloid, and other CSF markers of neurodegeneration.

METHODS

The authors studied copper, ceruloplasmin, total peroxide, and antioxidants levels (TRAP) in serum; beta-amyloid in plasma; and copper, beta-amyloid, h-tau, and P-tau in the CSF of 28 patients with AD and 25 healthy controls, in relation to clinical status.

RESULTS

Serum copper (p < 0.0001), peroxides (p = 0.002), a copper fraction unexplained by ceruloplasmin (p < 0.0001), and CSF h-tau (p = 0.001) were increased in AD, whereas serum TRAP (p = 0.03) and CSF beta-amyloid were decreased (p < 0.0001). Plasma beta-amyloid increased with age in healthy controls (r = 0.6; p = 0.05). CSF markers of AD correlated with serum copper variables. CSF copper was partially dependent on the serum copper fraction unexplained by ceruloplasmin (t = 2.2, p = 0.04). CSF beta-amyloid seemed to be related to serum copper (r = -0.46; p = 0.002). Mini-Mental Status Examination scores correlated positively with beta-amyloid (r = 0.46, p = 0.002) and inversely with copper unexplained by ceruloplasmin (r = -0.45, p = 0.003).

CONCLUSIONS

The authors' results confirm the existence of changes in copper component distribution, particularly the copper fraction unexplained by ceruloplasmin and support the hypothesis of a beta-amyloid and copper connection in Alzheimer disease.

Fetal auditory responses to external sounds and mother's heart beat: detection improved by Independent Component Analysis

In this paper, we present a magnetoencephalographic study of the fetal auditory response to external stimuli and to the sound of the mother's heartbeat. We describe how an ad hoc functional selection procedure allowed us to isolate the sources in the fetal brain responding to sounds only, after the application to the recorded data of a standard Independent Component Analysis algorithm. In our experiment, acoustic stimuli were delivered to twelve healthy women with uncomplicated pregnancies at a time between 36 and 40 weeks gestational age, with their fetuses in breech presentation. Ultrasound images allowed determination of the region over the women's abdomen nearest to the fetal head, over which both the acoustic stimulator and the MEG sensors were subsequently placed. In 8 out of the 12 cases, our analysis provided consistent evidence of a fetal response both to the mother's heartbeat and to the external auditory stimulation; both were characterized by a clear prominent component at around 200 ms latency, which is widely accepted as the marker of the fetal response to auditory stimuli.

Minimum local anaesthetic dose (MLAD) of intrathecal levobupivacaine and ropivacaine for Caesarean section*

We determined the minimum local anaesthetic dose (MLAD) of spinal levobupivacaine and ropivacaine for Caesarean section. Ninety women were randomly allocated to two groups and received 3 ml of study solution by a combined spinal/epidural technique. The initial dose was 12 mg for levobupivacaine and 17 mg for ropivacaine groups. To be considered effective, a test solution had to achieve a visual analogue pain score (VAPS) of 30 mm or less at skin incision, uterine incision, birth, peritoneal closure, and at the end of surgery. Effective or ineffective responses determined, respectively, a 0.3 mg decrease or increase of the same drug for the next patient in the same group, using up-down sequential allocation. The MLAD of levobupivacaine was 10.58 mg (CI 95%: 10.08-11.09) and the MLAD of ropivacaine 14.22 mg (CI 95%: 13.67-14.77), using the Dixon and Massey formula. The potency ratio between spinal levobupivacaine and spinal ropivacaine was 1.34.

Longitudinal profiling of urinary steroids by gas chromatography/combustion/isotope ratio mass spectrometry: Diet change may result in carbon isotopic variations

Longitudinal profiling of urinary steroids was investigated by using a gas chromatography/combustion/isotope ratio mass spectrometry (GC/C/IRMS) method. The carbon isotope ratio of three urinary testosterone (T) metabolites: androsterone, etiocholanolone, 5beta-androstane-3alpha,17beta-diol (5beta-androstanediol) together with 16(5alpha)-androsten-3alpha-ol (androstenol) and 5beta-pregnane-3alpha,20alpha-diol (5beta-pregnanediol) were measured in urine samples collected from three top-level athletes over 2 years. Throughout the study, the subjects were living in Switzerland and were residing every year for a month or two in an African country. (13)C-enrichment larger than 2.5 per thousand was observed for one subject after a 2-month stay in Africa. Our findings reveal that (13)C-enrichment caused by a diet change might be reduced if the stay in Africa was shorter or if the urine sample was not collected within the days after return to Switzerland. The steroids of interest in each sample did not show significant isotopic fractionation that could lead to false positive results in anti-doping testing. In contrast to the results obtained with the carbon isotopic ratio, profiling of urinary testosterone/epitestosterone (T/E) ratios was found to be unaffected by a diet change.