MGOF: Stata module to perform goodness-of-fit tests for multinomial data

mgof computes goodness-of-fit tests for the distribution of a discrete (categorical, multinomial) variable. The default is to perform classical large sample chi-squared approximation tests based on Pearson's X2 statistic and the log likelihood ratio (G2) statistic or a statistic from the Cressie-Read family. Alternatively, mgof computes exact tests using Monte Carlo methods or exhaustive enumeration. A Kolmogorov-Smirnov test for discrete data is also provided. The moremata package, also available from SSC, is required.

Multinomial goodness-of-fit: large sample tests with survey design correction and exact tests for small samples

A new Stata command called -mgof- is introduced. The command is used to compute distributional tests for discrete (categorical, multinomial) variables. Apart from classic large sample $\chi^2$-approximation tests based on Pearson's $X^2$, the likelihood ratio, or any other statistic from the power-divergence family (Cressie and Read 1984), large sample tests for complex survey designs and exact tests for small samples are supported. The complex survey correction is based on the approach by Rao and Scott (1981) and parallels the survey design correction used for independence tests in -svy:tabulate-. The exact tests are computed using Monte Carlo methods or exhaustive enumeration. An exact Kolmogorov-Smirnov test for discrete data is also provided.

Multinomial goodness-of-fit: large sample tests with survey design correction and exact tests for small samples

I introduce the new mgof command to compute distributional tests for discrete (categorical, multinomial) variables. The command supports largesample tests for complex survey designs and exact tests for small samples as well as classic large-sample x2-approximation tests based on Pearson’s X2, the likelihood ratio, or any other statistic from the power-divergence family (Cressie and Read, 1984, Journal of the Royal Statistical Society, Series B (Methodological) 46: 440–464). The complex survey correction is based on the approach by Rao and Scott (1981, Journal of the American Statistical Association 76: 221–230) and parallels the survey design correction used for independence tests in svy: tabulate. mgof computes the exact tests by using Monte Carlo methods or exhaustive enumeration. mgof also provides an exact one-sample Kolmogorov–Smirnov test for discrete data.

On Graphically Checking Goodness-of-fit of Binary Logistic Regression Models

OBJECTIVES: This paper is concerned with checking goodness-of-fit of binary logistic regression models. For the practitioners of data analysis, the broad classes of procedures for checking goodness-of-fit available in the literature are described. The challenges of model checking in the context of binary logistic regression are reviewed. As a viable solution, a simple graphical procedure for checking goodness-of-fit is proposed. METHODS: The graphical procedure proposed relies on pieces of information available from any logistic analysis; the focus is on combining and presenting these in an informative way. RESULTS: The information gained using this approach is presented with three examples. In the discussion, the proposed method is put into context and compared with other graphical procedures for checking goodness-of-fit of binary logistic models available in the literature. CONCLUSION: A simple graphical method can significantly improve the understanding of any logistic regression analysis and help to prevent faulty conclusions.

In vitro precision of fit of computer-aided design and computer-aided manufacturing titanium and zirconium dioxide bars

OBJECTIVES Optical scanners combined with computer-aided design and computer-aided manufacturing (CAD/CAM) technology provide high accuracy in the fabrication of titanium (TIT) and zirconium dioxide (ZrO) bars. The aim of this study was to compare the precision of fit of CAD/CAM TIT bars produced with a photogrammetric and a laser scanner. METHODS Twenty rigid CAD/CAM bars were fabricated on one single edentulous master cast with 6 implants in the positions of the second premolars, canines and central incisors. A photogrammetric scanner (P) provided digitized data for TIT-P (n=5) while a laser scanner (L) was used for TIT-L (n=5). The control groups consisted of soldered gold bars (gold, n=5) and ZrO-P with similar bar design. Median vertical distance between implant and bar platforms from non-tightened implants (one-screw test) was calculated from mesial, buccal and distal scanning electron microscope measurements. RESULTS Vertical microgaps were not significantly different between TIT-P (median 16μm; 95% CI 10-27μm) and TIT-L (25μm; 13-32μm). Gold (49μm; 12-69μm) had higher values than TIT-P (p=0.001) and TIT-L (p=0.008), while ZrO-P (35μm; 17-55μm) exhibited higher values than TIT-P (p=0.023). Misfit values increased in all groups from implant position 23 (3 units) to 15 (10 units), while in gold and TIT-P values decreased from implant 11 toward the most distal implant 15. SIGNIFICANCE CAD/CAM titanium bars showed high precision of fit using photogrammetric and laser scanners. In comparison, the misfit of ZrO bars (CAM/CAM, photogrammetric scanner) and soldered gold bars was statistically higher but values were clinically acceptable.

In vitro precision of fit of computer-aided designed and computer-aided manufactured titanium screw-retained fixed dental prostheses before and after ceramic veneering

OBJECTIVE To compare the precision of fit of full-arch implant-supported screw-retained computer-aided designed and computer-aided manufactured (CAD/CAM) titanium-fixed dental prostheses (FDP) before and after veneering. The null-hypothesis was that there is no difference in vertical microgap values between pure titanium frameworks and FDPs after porcelain firing. MATERIALS AND METHODS Five CAD/CAM titanium grade IV frameworks for a screw-retained 10-unit implant-supported reconstruction on six implants (FDI tooth positions 15, 13, 11, 21, 23, 25) were fabricated after digitizing the implant platforms and the cuspid-supporting framework resin pattern with a laser scanner (CARES(®) Scan CS2; Institut Straumann AG, Basel, Switzerland). A bonder, an opaquer, three layers of porcelain, and one layer of glaze were applied (Vita Titankeramik) and fired according to the manufacturer's preheating and fire cycle instructions at 400-800°C. The one-screw test (implant 25 screw-retained) was applied before and after veneering of the FDPs to assess the vertical microgap between implant and framework platform with a scanning electron microscope. The mean microgap was calculated from interproximal and buccal values. Statistical comparison was performed with non-parametric tests. RESULTS All vertical microgaps were clinically acceptable with values <90 μm. No statistically significant pairwise difference (P = 0.98) was observed between the relative effects of vertical microgap of unveneered (median 19 μm; 95% CI 13-35 μm) and veneered FDPs (20 μm; 13-31 μm), providing support for the null-hypothesis. Analysis within the groups showed significantly different values between the five implants of the FDPs before (P = 0.044) and after veneering (P = 0.020), while a monotonous trend of increasing values from implant 23 (closest position to screw-retained implant 25) to 15 (most distant implant) could not be observed (P = 0.169, P = 0.270). CONCLUSIONS Full-arch CAD/CAM titanium screw-retained frameworks have a high accuracy. Porcelain firing procedure had no impact on the precision of fit of the final FDPs. All implant microgap measurements of each FDP showed clinically acceptable vertical misfit values before and after veneering. Thus, the results do not only show accurate performance of the milling and firing but show also a reproducible scanning and designing process.

Precision of fit of implant-supported screw-retained 10-unit computer-aided-designed and computer-aided-manufactured frameworks made from zirconium dioxide and titanium: an in vitro study

OBJECTIVE To analyze the precision of fit of implant-supported screw-retained computer-aided-designed and computer-aided-manufactured (CAD/CAM) zirconium dioxide (ZrO) frameworks. MATERIALS AND METHODS Computer-aided-designed and computer-aided-manufactured ZrO frameworks (NobelProcera) for a screw-retained 10-unit implant-supported reconstruction on six implants (FDI positions 15, 13, 11, 21, 23, 25) were fabricated using a laser (ZrO-L, N = 6) and a mechanical scanner (ZrO-M, N = 5) for digitizing the implant platform and the cuspid-supporting framework resin pattern. Laser-scanned CAD/CAM titanium (TIT-L, N = 6) and cast CoCrW-alloy frameworks (Cast, N = 5) fabricated on the same model and designed similar to the ZrO frameworks were the control. The one-screw test (implant 25 screw-retained) was applied to assess the vertical microgap between implant and framework platform with a scanning electron microscope. The mean microgap was calculated from approximal and buccal values. Statistical comparison was performed with non-parametric tests. RESULTS No statistically significant pairwise difference was observed between the relative effects of vertical microgap between ZrO-L (median 14 μm; 95% CI 10-26 μm), ZrO-M (18 μm; 12-27 μm) and TIT-L (15 μm; 6-18 μm), whereas the values of Cast (236 μm; 181-301 μm) were significantly higher (P < 0.001) than the three CAD/CAM groups. A monotonous trend of increasing values from implant 23 to 15 was observed in all groups (ZrO-L, ZrO-M and Cast P < 0.001, TIT-L P = 0.044). CONCLUSIONS Optical and tactile scanners with CAD/CAM technology allow for the fabrication of highly accurate long-span screw-retained ZrO implant-reconstructions. Titanium frameworks showed the most consistent precision. Fit of the cast alloy frameworks was clinically inacceptable.

DIGDIS: Stata module to analyze the distribution of digits

digdis tabulates the distribution of digits of the specified variables, performs goodness-of-fit tests against a reference distribution and, optionally, graphs the distributions. The default is to tabulate the first (nonzero) digit and to test against Benford's law. The moremata package and the mgof package, also available from SSC, are required.

EEG correlation and power during maintenance of wakefulness test after sleep-deprivation

To investigate whether there are any objective EEG characteristics that change significantly between specific time periods during maintenance of wakefulness test (MWT) and whether such changes are associated with the ability to appropriately communicate sleepiness.

Multicentre evaluation of a new point-of-care test for the determination of NT-proBNP in whole blood

BACKGROUND: The Roche CARDIAC proBNP point-of-care (POC) test is the first test intended for the quantitative determination of N-terminal pro-brain natriuretic peptide (NT-proBNP) in whole blood as an aid in the diagnosis of suspected congestive heart failure, in the monitoring of patients with compensated left-ventricular dysfunction and in the risk stratification of patients with acute coronary syndromes. METHODS: A multicentre evaluation was carried out to assess the analytical performance of the POC NT-proBNP test at seven different sites. RESULTS: The majority of all coefficients of variation (CVs) obtained for within-series imprecision using native blood samples was below 10% for both 52 samples measured ten times and for 674 samples measured in duplicate. Using quality control material, the majority of CV values for day-to-day imprecision were below 14% for the low control level and below 13% for the high control level. In method comparisons for four lots of the POC NT-proBNP test with the laboratory reference method (Elecsys proBNP), the slope ranged from 0.93 to 1.10 and the intercept ranged from 1.8 to 6.9. The bias found between venous and arterial blood with the POC NT-proBNP method was < or =5%. All four lots of the POC NT-proBNP test investigated showed excellent agreement, with mean differences of between -5% and +4%. No significant interference was observed with lipaemic blood (triglyceride concentrations up to 6.3 mmol/L), icteric blood (bilirubin concentrations up to 582 micromol/L), haemolytic blood (haemoglobin concentrations up to 62 mg/L), biotin (up to 10 mg/L), rheumatoid factor (up to 42 IU/mL), or with 50 out of 52 standard or cardiological drugs in therapeutic concentrations. With bisoprolol and BNP, somewhat higher bias in the low NT-proBNP concentration range (<175 ng/L) was found. Haematocrit values between 28% and 58% had no influence on the test result. Interference may be caused by human anti-mouse antibodies (HAMA) types 1 and 2. No significant influence on the results with POC NT-proBNP was found using volumes of 140-165 muL. High NT-proBNP concentrations above the measuring range of the POC NT-proBNP test did not lead to false low results due to a potential high-dose hook effect. CONCLUSIONS: The POC NT-proBNP test showed good analytical performance and excellent agreement with the laboratory method. The POC NT-proBNP assay is therefore suitable in the POC setting.

Prediction of first coronary events with the Framingham score: a systematic review

BACKGROUND: Uncertainty exists about the performance of the Framingham risk score when applied in different populations. OBJECTIVE: We assessed calibration of the Framingham risk score (ie, relationship between predicted and observed coronary event rates) in US and non-US populations free of cardiovascular disease. METHODS: We reviewed studies that evaluated the performance of the Framingham risk score to predict first coronary events in a validation cohort, as identified by Medline, EMBASE, BIOSIS, and Cochrane library searches (through August 2005). Two reviewers independently assessed 1496 studies for eligibility, extracted data, and performed quality assessment using predefined forms. RESULTS: We included 25 validation cohorts of different population groups (n = 128,000) in our main analysis. Calibration varied over a wide range from under- to overprediction of absolute risk by factors of 0.57 to 2.7. Risk prediction for 7 cohorts (n = 18658) from the United States, Australia, and New Zealand was well calibrated (corresponding figures: 0.87-1.08; for the 5 biggest cohorts). The estimated population risks for first coronary events were strongly associated (goodness of fit: R2 = 0.84) and in good agreement with observed risks (coefficient for predicted risk: beta = 0.84; 95% CI 0.41-1.26). In 18 European cohorts (n = 109499), the corresponding figures indicated close association (R2 = 0.72) but substantial overprediction (beta = 0.58, 95% CI 0.39-0.77). The risk score was well calibrated on the intercept for both population clusters. CONCLUSION: The Framingham score is well calibrated to predict first coronary events in populations from the United States, Australia, and New Zealand. Overestimation of absolute risk in European cohorts requires recalibration procedures.