Confidence Interval For Risk Difference Research Articles

Continuity corrected score confidence interval for the difference in proportions in paired data

For paired binary data, the hybrid method and the score method are often recommended for use to calculate the confidence interval for risk difference. These asymptotic intervals do not control the coverage probability. We propose to develop a new score interval with continuity correction to further improve the performance of the existing intervals. The traditional correction value may be too large which leads to a wide interval. For that reason, we propose three different correction values to identify the optimal correction interval with balanced coverage probability and interval width. From simulation studies, we find that a small correction value for the score interval has good performance. In addition, we derive the non-iterative solutions for the developed continuity correction score intervals.

Accurate interval estimation for the risk difference in an incomplete correlated 2 × 2 table: Calf immunity analysis.

Interval estimation with accurate coverage for risk difference (RD) in a correlated 2 × 2 table with structural zero is a fundamental and important problem in biostatistics. The score test-based and Bayesian tail-based confidence intervals (CIs) have good coverage performance among the existing methods. However, as approximation approaches, they have coverage probabilities lower than the nominal confidence level for finite and moderate sample sizes. In this paper, we propose three new CIs for RD based on the fiducial, inferential model (IM) and modified IM (MIM) methods. The IM interval is proven to be valid. Moreover, simulation studies show that the CIs of fiducial and MIM methods can guarantee the preset coverage rate even for small sample sizes. More importantly, in terms of coverage probability and expected length, the MIM interval outperforms other intervals. Finally, a real example illustrates the application of the proposed methods.

Exact-corrected confidence interval for risk difference in noninferiority binomial trials.

A novel confidence interval estimator is proposed for the risk difference in noninferiority binomial trials. The proposed confidence interval, which is dependent on the prespecified noninferiority margin, is consistent with an exact unconditional test that preserves the type-I error and has improved power, particularly for smaller sample sizes, compared to the confidence interval by Chan and Zhang. The improved performance of the proposed confidence interval is theoretically justified and demonstrated with simulations and examples. An R package is also distributed that implements the proposed methods along with other confidence intervalestimators.

Implementation of Miettinen-Nurminen score method with or without stratification in R

Analysis of a 2 × 2 table for clinical data involves computing the point estimate and confidence interval for risk difference, relative risk, or odds ratio. While point estimates of these comparative parameters are uniquely defined, several statistical methods have been proposed to estimate the confidence interval for each parameter. The Miettinen-Nurminen (MN) score method is expected to be used increasingly over traditional interval estimation methods. The MN score method has not been previously implemented in R software for data with stratification. There is a need for a comprehensive software implementation of the MN score method. This article describes the implementation of the MN score method in the sasLM R software package. To demonstrate the usage of the sasLM functions introduced, recently published clinical data are provided as examples.

Noninferiority of Microhook to Trabectome: Trabectome versus Ab Interno Microhook Trabeculotomy Comparative Study (Tram Trac Study)

To elucidate the noninferiority of ab interno microhook trabeculotomy (μTLO) using a recently developed reusable stainless spatula-type microhook device to incise the trabecular meshwork to Trabectome (Neomeix, Inc) surgery in terms of the 1-year postoperative outcomes of Japanese patients with glaucoma by means of propensity score analyses. Multicenter, retrospective cohort study. We enrolled 553 and 392 patients who underwent Trabectome surgery and μTLO, respectively, between January 2014 and March 2020 at 10 facilities. Logistic regression analysis was conducted to calculate the propensity score, which indicates the likelihood of treatment assignment (Trabectome or μTLO). We set the following factors as outcome-related covariates: age, sex, facility, glaucoma disease types, preoperative intraocular pressure (IOP), glaucoma drug score, mean deviation of Humphrey visual field test results, antithrombotic drug use, the presence or absence of combined cataract surgery, and incision range of the trabecular meshwork (1 or 2 quadrants). We analyzed 4 different methods (matching, inverse probability of treatment weighting [IPTW], stratification, and regression adjustment) using the propensity score. We set 15% as the noninferiority margin based on previous Trabectome meta-analysis results. The primary outcome was surgical success at 1 year after surgery. We defined surgical success as satisfying all 3 criteria: (1) IOP within 5 to 21 mmHg, (2) IOP reduction of 20% or more from preoperative IOP, and (3) no additional glaucoma surgery. The 95% confidence interval of risk difference of surgical failure in μTLO in reference to Trabectome surgery was-12.1% to+9.5% in matching,-12.7% to+11.1% in IPTW,-12.2 to+7.0 in stratification, and-9.7% to+8.1% in regression adjustment, all of which fell within the predetermined noninferiority margin of 15%. Surgical success of μTLO at 1 year after was not inferior to that of Trabectome surgery.

Confidence interval of risk difference by different statistical methods and its impact on the study conclusion in antibiotic non-inferiority trials

BackgroundNumerous statistical methods can be used to calculate the confidence interval (CI) of risk differences. There is consensus in previous literature that the Wald method should be discouraged. We compared five statistical methods for estimating the CI of risk difference in terms of CI width and study conclusion in antibiotic non-inferiority trials.MethodsIn a secondary analysis of a systematic review, we included non-inferiority trials that compared different antibiotic regimens, reported risk differences for the primary outcome, and described the number of successes and/or failures as well as patients in each arm. For each study, we re-calculated the risk difference CI using the Wald, Agresti-Caffo, Newcombe, Miettinen-Nurminen, and skewness-corrected asymptotic score (SCAS) methods. The CIs by different statistical methods were compared in terms of CI width and conclusion on non-inferiority. A wider CI was considered to be more conservative.ResultsThe analysis included 224 comparisons from 213 studies. The statistical method used to calculate CI was not reported in 134 (59.8%) cases. The median (interquartile range IQR) for CI width by Wald, Agresti-Caffo, Newcombe, Miettinen-Nurminen, and SCAS methods was 13.0% (10.8%, 17.4%), 13.3% (10.9%, 18.5%), 13.6% (11.1%, 18.9%), 13.6% (11.1% and 19.0%), and 13.4% (11.1%, 18.9%), respectively. In 216 comparisons that reported a non-inferiority margin, the conclusion on non-inferiority was the same across the five statistical methods in 211 (97.7%) cases. The differences in CI width were more in trials with a sample size of 100 or less in each group and treatment success rate above 90%. Of the 18 trials in this subgroup with a specified non-inferiority margin, non-inferiority was shown in 17 (94.4%), 16 (88.9%), 14 (77.8%), 14 (77.8%), and 15 (83.3%) cases based on CI by Wald, Agresti-Caffo, Newcombe, Miettinen-Nurminen, and SCAS methods, respectively.ConclusionsThe statistical method used to calculate CI was not reported in the majority of antibiotic non-inferiority trials. Different statistical methods for CI resulted in different conclusions on non-inferiority in 2.3% cases. The differences in CI widths were highest in trials with a sample size of 100 or less in each group and a treatment success rate above 90%.Trial registrationPROSPERO CRD42020165040. April 28, 2020.

Relationship between surgical R0 resectability and findings of peripancreatic vascular invasion on CT imaging after neoadjuvant S-1 and concurrent radiotherapy in patients with borderline resectable pancreatic cancer

BackgroundBorderline resectable pancreatic cancer (BRPC) is frequently associated with positive surgical margins and a poor prognosis because the tumor is in contact with major vessels. This study evaluated the relationship between the margin-negative (R0) resection rate and findings indicating peripancreatic vascular invasion on multidetector computed tomography (MDCT) imaging after neoadjuvant chemoradiotherapy (NACRT) in patients with BRPC.MethodsTwenty-nine BRPC patients who underwent laparotomy after neoadjuvant S-1 with concurrent radiotherapy were studied retrospectively. Peripancreatic major vessel invasion was evaluated based on the length of tumor-vessel contact on MDCT. The R0 resection rates were compared between the progression of vascular invasion (PVI) group and the non-progression of vascular invasion (NVI) group.ResultsThere were 3 patients with partial responses (10%), 25 with stable disease (86%), and 1 with progressive disease (3%) according to the RECISTv1.1 criteria. Regarding vascular invasion, 9 patients (31%) were classified as having PVI, and 20 patients (69%) were classified as having NVI. Of the 29 patients, 27 (93%) received an R0 resection, and all the PVI patients received an R0 resection (9/9; R0 resection rate = 100%) while 90% (18/20) of the NVI patients underwent an R0 resection. The exact 95% confidence interval of risk difference between those R0 resection rates was − 10.0% [− 31.7–20.4%].ConclusionsPatients with BRPC after NACRT achieved high R0 resection rates regardless of the vascular invasion status. BRPC patients can undergo R0 resections unless progressive disease is observed after NACRT.Trial registrationUMIN-CTR, UMIN000009172. Registered 23 October 2012

Prevalence of SARS-CoV-2 infection in previously undiagnosed health care workers in New Jersey, at the onset of the U.S. COVID-19 pandemic

BackgroundHealthcare workers (HCW) are presumed to be at increased risk of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection due to occupational exposure to infected patients. However, there has been little epidemiological research to assess these risks.MethodsWe conducted a prospective cohort study of HCW (n = 546) and non-healthcare workers (NHCW; n = 283) with no known prior SARS-CoV-2 infection who were recruited from a large U.S. university and two affiliated university hospitals. In this cross-sectional analysis of data collected at baseline, we examined SARS-CoV-2 infection status (as determined by presence of SARS-CoV-2 RNA in oropharyngeal swabs) by healthcare worker status and role.ResultsAt baseline, 41 (5.0%) of the participants tested positive for SARS-CoV-2 infection, of whom 14 (34.2%) reported symptoms. The prevalence of SARS-CoV-2 infection was higher among HCW (7.3%) than in NHCW (0.4%), representing a 7.0% greater absolute risk (95% confidence interval for risk difference 4.7, 9.3%). The majority of infected HCW (62.5%) were nurses. Positive tests increased across the two weeks of cohort recruitment in line with rising confirmed cases in the hospitals and surrounding counties.ConclusionsOverall, our results demonstrate that HCW had a higher prevalence of SARS-CoV-2 infection than NHCW. Continued follow-up of this cohort will enable us to monitor infection rates and examine risk factors for transmission.

Performance of matching methods in studies of rare diseases: a simulation study.

Matching is a common method of adjusting for confounding in observational studies. Studies in rare diseases usually include small numbers of exposed subjects, but the performance of matching methods in such cases has not been evaluated thoroughly. In this study, we compare the performance of several matching methods when number of exposed subjects is small. We used Monte Carlo simulations to compare the following methods: Propensity score matching (PSM) with greedy or optimal algorithm, Mahalanobis distance matching, and mixture of PSM and exact matching. We performed the comparisons in datasets with six continuous and six binary variables, with varying effect size on group assignment and outcome. In each case, there were 1,500 unexposed subjects and a varying number of exposed: N = 25, 50, 100, 150, 200, 250, or 300. The probability of outcome in unexposed subjects was set to 5% (rare), 20% (common), or 50% (frequent). We compared the methods based on the bias of estimate of risk difference, coverage of 95% confidence intervals for risk difference, and balance of covariates. We observed a difference in performance of matching methods in very small samples (N = 25-50) and in moderately small samples (N = 100-300). Our study showed that PSM performs better than other matching methods when number of exposed subjects is small, but the matching algorithm and the matching ratio should be considered carefully. We recommend using PSM with optimal algorithm and one-to-five matching ratio in very small samples, and PSM matching with any algorithm and one-to-one matching in moderately small samples.

Efficacy and safety of ABP 798 compared with rituximab: Results from the comparative clinical study in patients with non-Hodgkin’s.

8044 Background: ABP 798* is being developed as a biosimilar to rituximab, a CD20-directed cytolytic antibody. A randomized, double-blind, active-controlled study compared the efficacy, safety, and immunogenicity of ABP 798 with rituximab reference product (RP) in subjects with CD20-positive NHL; results of the final analysis are presented here. Methods: Adult subjects with grade 1, 2, or 3a follicular B-cell NHL and low tumor burden were randomized to receive intravenous ABP 798 or RP (375 mg/m2) once weekly for 4 weeks, then at weeks 12 and 20. The primary endpoint was risk difference (RD) of overall response rate (ORR) by week 28. Secondary endpoints included RD of ORR at week 12, pharmacokinetics, pharmacodynamics, safety, and immunogenicity. Results: 254/256 randomized subjects were treated with at least one infusion of ABP 798 (n = 128) or RP (n = 126); ORR by week 28, based on independent central blinded assessment of the modified full analysis set, was comparable between the ABP 798 and RP groups (78% vs. 70%, respectively). The 2-sided 90% confidence interval of RD of ORR (-1.4%; 16.8%) was within the pre-specified margin (-15%; 35.5%) thereby establishing clinical equivalence between ABP 798 and RP. This result was supported by analyses of the secondary efficacy endpoint of RD of ORR at week 12. In the two groups, the geometric least squares means for serum concentrations over time (e.g., week 12, pre-dose: ABP 798, 21.89 vs. RP, 20.57; week 12 post-dose: ABP 798, 201.30 vs. RP, 203.52) and the extent of B-cell depletion from day 1 to day 8 (ABP 798, 98.3% vs. RP, 98.3%) were similar. Frequency, type, and severity of adverse events (AEs) were comparable between ABP 798 and RP groups; grade ≥3 AEs were reported in 10.9% and 10.3% of subjects and serious AEs in 3.9% and 4.0%, respectively. Most common AEs were headache, fatigue, and nausea; the most common AE of interest was infusion reactions. No new or unexpected safety signals were observed. Binding and neutralizing anti-drug antibodies were comparable between groups. Conclusions: Results of this study demonstrated clinical similarity between ABP 798 and rituximab RP in subjects with CD20-positive NHL. *At the time of this submission, ABP 798 had not been approved by the FDA or any relevant regulatory agency and the indications are yet undetermined. Please consult ABP 798’s later approved label in the relevant country for information regarding the approved uses for ABP 798.

Accurate confidence intervals for risk difference in meta-analysis with rare events

BackgroundMeta-analysis provides a useful statistical tool to effectively estimate treatment effect from multiple studies. When the outcome is binary and it is rare (e.g., safety data in clinical trials), the traditionally used methods may have unsatisfactory performance.MethodsWe propose using importance sampling to compute confidence intervals for risk difference in meta-analysis with rare events. The proposed intervals are not exact, but they often have the coverage probabilities close to the nominal level. We compare the proposed accurate intervals with the existing intervals from the fixed- or random-effects models and the interval by Tian et al. (2009).ResultsWe conduct extensive simulation studies to compare them with regards to coverage probability and average length, when data are simulated under the homogeneity or heterogeneity assumption of study effects.ConclusionsThe proposed accurate interval based on the random-effects model for sample space ordering generally has satisfactory performance under the heterogeneity assumption, while the traditionally used interval based on the fixed-effects model works well when the studies are homogeneous.

Abstract OT1-03-04: The importance of central pathology pCR review in an international multicenter neoadjuvant study in HER2 positive early breast cancer - Results of the ABP 980 LILAC trial

Abstract Pathological complete response (pCR) after neoadjuvant therapy is an accepted surrogate endpoint for breast cancer registration trials and is a strong prognostic biomarker especially in high-risk tumor biology. The LILAC study, a comparative trial for the trastuzumab biosimilar ABP 980, was the first international multicenter neoadjuvant breast cancer trial using local and central pathology review for pCR. Here we report concordance between local and central review for the most commonly used pCR definitions. This study included 725 patients with early HER2-positive breast cancer, comparing neoadjuvant ABP 980 and trastuzumab reference product (RP) in combination with anthracycline and taxane-containing chemotherapy. Equivalence of both arms was defined at a significance level of 0.05 by comparing the 2-sided 90% confidence interval (CI) for the risk difference (RD) of pCR between ABP 980 and RP within a predefined margin (-13%, 13%) using definition 2 as described below. pCR was evaluated in local laboratories and in a central laboratory blinded to the local result. The purpose of the independent, central review was to reduce the inter-pathologist variability and provide a more consistent assessment. Central review was performed by 2 independent pathologists; in cases of discordance the specimen was reviewed and decided by a third pathologist (adjudicator). The pCR definition in the primary analysis was no invasive breast cancer in breast and axillary nodes, regardless of ductal carcinoma in situ (DCIS). No invasive tumor in the breast was pCR definition 1, no invasive breast cancer in breast and axillary nodes regardless of DCIS was pCR definition 2, and no invasive breast cancer in breast and axillary nodes and absence of DCIS was pCR definition 3. The RDs and CIs were evaluated for local and central pCR review for all 3 pCR definitions. The results of the neoadjuvant LILAC study have been reported previously, demonstrating no clinically meaningful differences between ABP 980 and RP in HER2-positive early breast cancer. Of the 696 subjects who underwent surgery and had specimens evaluable for pCR review by the local pathologists, 669 specimens were available for central review. Rates of pCR in the local review were 51.1% (ABP 980) and 45% (RP) for definition 1, 48% (ABP 980) and 40.5% (RP) for definition 2, and 37.7% (ABP 980) and 29.6% (RP) for definition 3. The corresponding pCR rates in the central review were 51.3% and 47.3% for definition 1, 47.8% and 41.8% for definition 2, and 29.9% and 26.1% for definition 3. The CI for RD was within the predefined margins for pCR definition 1 in local and central review, whereas it exceeded the prespecified margin in local review for pCR definitions 2 and 3. In the central review the CI was within the margin for the pCR definitions 2 and 3, with the difference between local and central pathology review 1.6% for pCR definition 2 and 4.4% for pCR definition 3, respectively. RDs (90% CI) for all 3 pCR definitions are shown in the Table. The LILAC study demonstrated that a central pathology review for pCR is feasible in a large, international, multicenter neoadjuvant breast cancer trial. Our results show that concordance between local and central pathology can decrease with increasing complexity of the pCR definition. Based on the results, it may be justified to include a central pathology review in neoadjuvant multicenter breast cancer trials with strict pCR definitions. Citation Format: Hanz-Christian Kolberg, Marco Colleoni, Georgia Savva Demetriou, Patricia Santi, Hans Tesch, Yasuhiro Fujiwara, Vladimir Hanes. The importance of central pathology pCR review in an international multicenter neoadjuvant study in HER2 positive early breast cancer - Results of the ABP 980 LILAC trial [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr OT1-03-04.

Anti-Ebola therapy for patients with Ebola virus disease: a systematic review

BackgroundManagement of Ebola virus disease (EVD) has historically focused on infection prevention, case detection and supportive care. Several specific anti-Ebola therapies have been investigated, including during the 2014–2016 West African outbreak. Our objective was to conduct a systematic review of the effect of anti-Ebola virus therapies on clinical outcomes to guide their potential use and future evaluation.MethodsWe searched PubMed, EMBASE, Global Health, Cochrane Library, African Index Medicus, WHOLIS (inception-9 April 2018), and trial registries for observational studies or clinical trials, in any language, that enrolled patients with confirmed EVD who received therapy targeting Ebola virus and reported on mortality, symptom duration, or adverse effects.ResultsFrom 11,257 citations and registered trials, we reviewed 55 full-text citations, of which 35 met eligibility criteria (1 randomized clinical trial (RCT), 8 non-randomized comparative studies, 9 case series and 17 case reports) and collectively examined 21 anti-Ebola virus agents. The 31 studies performed during the West African outbreak reported on 4.8% (1377/28616) of all patients with Ebola. The only RCT enrolled 72 patients (0.25% of all patients with Ebola) and compared the monoclonal antibody ZMapp vs. standard care (mortality, 22% vs. 37%; 95% confidence interval for risk difference, − 36 to 7%). Studies of convalescent plasma, interferon-β-1a, favipiravir, brincidofovir, artesunate-amodiaquine and TKM-130803 were associated with at least moderate risk of bias.ConclusionsResearch evaluating anti-Ebola virus agents has reached very few patients with EVD, and inferences are limited by non-randomized study designs. ZMapp has the most promising treatment signal.

Dose‐associated pulmonary complication rates after fresh frozen plasma administration for warfarin reversal

ESSENTIALS: Fresh frozen plasma (FFP) may be associated with a dose-based risk of pulmonary complications. Patients received FFP for warfarin reversal at a large academic hospital over a 3-year period. Almost 20% developed pulmonary complications, and the risk was highest after > 3 units of FFP. The risk of pulmonary complications remained significant in multivariable analysis. Fresh frozen plasma (FFP) is often administered to reverse warfarin anticoagulation. Administration has been associated with pulmonary complications, but it is unclear whether this risk is dose-related. Aims We sought to characterize the incidence and dose relationship of pulmonary complications, including transfusion-associated circulatory overload (TACO) and transfusion-related acute lung injury (TRALI), after FFP administration for warfarin reversal. We performed a structured retrospective review of patients who received FFP for warfarin reversal in the emergency department (ED) of an academic tertiary-care hospital over a 3-year period. Logistic regression was used to explore the relationship between FFP dose and risk of pulmonary events. Two hundred and fifty-one patients met the inclusion criteria. Overall, 49 patients (20%) developed pulmonary complications, including 30 (12%) with TACO, two (1%) with TRALI, and 17 (7%) with pulmonary edema not meeting the criteria for TACO. Pulmonary complications were significantly more frequent in those who received > 3 units of FFP (34.0% versus 15.6%, 95% confidence interval for risk difference 7.9%-8.9%). After stratification by subtype of complication, only the risk of TACO was statistically significant (28.3% versus 7.6%, 95% confidence interval for risk difference 8.2%-16.6%). In multivariable analysis controlling for age, sex, initial systolic blood pressure, and intravenous fluids given in the ED, > 3 units of FFP remained a significant risk factor for pulmonary complications (odds ratio 2.49, 95% confidence interval 1.21-5.13). Almost 20% of patients who received FFP for warfarin reversal developed pulmonary complications, primarily TACO, and this risk increased with > 3 units of FFP. Clinicians should be aware of and prepared to manage these complications.

Biological drugs for the treatment of moderate-to-severe psoriasis by subcutaneous route: determining statistical equivalence according to evidence-based methods.

Although several evidence-based analyses have found that in patients with moderate-to-severe psoriasis, biological agents for subcutaneous use show no difference in efficacy (no proof of difference), there is limited evidence on whether or not there is also proof of no difference. The clinical material was the same as that reported in previous meta-analyses. Our methodology was a combination of meta-analysis and equivalence testing. The endpoint was the rate of Psoriasis Area and Severity Index (PASI) 75 achievement. The agents examined for equivalence testing included etanercept, high-dose etanercept, adalimumab, ustekinumab, and high-dose ustekinumab. The equivalence margin [±18 % as risk difference (RD)] was derived from statistical power information from the original trials; a more conservative margin at ±10 % was also tested. Our analysis involved 16 randomized trials including 8,257 patients. Ten head-to-head indirect comparisons were made and the respective values of RD were estimated. The 95 % confidence intervals for RDs remained within the margins of ±18 and ±10 % in six and two cases, respectively. A post hoc margin of about ±25 % was satisfied in all cases. All analyses assumed α = 2.5 %. Our study indicates that some of these five biological treatments can be equivalent, but results are strongly influenced by the margins adopted. Our findings can be helpful to develop local acquisition tenders on these drugs.

Likelihood-based confidence intervals for the risk difference of two-sample binary data with a fallible classifier and a gold standard

We develop likelihood-based confidence intervals for risk difference in two-sample misclassified binary data. Such data consist of two studies. The first study is the main study where individuals are classified by an inexpensive fallible classifier which may misclassify. The second study is a validation substudy where individuals are classified by using both the fallible classifier and an expensive gold standard which classifies perfectly. We propose and examine three likelihood-based confidence interval methods and conclude that the modified Wald method applied to small-number adjusted new data performs well and has nominal coverage probabilities.

Interval Estimation of Some Epidemiological Measures of Association

In epidemiological cohort studies, the probability of developing a disease for individuals in a treatment/intervention group is compared with that of a control group. The groups involve varying cluster sizes, and the binary responses within each cluster cannot be assumed independently. Three major measures of association used to report the efficacy of treatments or effectiveness of public health intervention programs in case of prospective studies are Risk Difference (RD), Risk Ratio (RR) and Relative Risk Difference (RED). The preference of one measure of association over the other in drawing statistical inference depends on design of study. Lui (2004) discusses a number of methods of constructing confidence intervals for each of these measures. Specifically, Lui (2004) discusses four methods for RD, four methods for RR and three methods for RED. For the construction of confidence intervals for RD, Paul and Zaihra (2008) compare the four methods discussed by Lui (2004), using extensive simulations with a method based on an estimator of the variance of a ratio estimator by Cochran (1977) and a method based on a sandwich estimator of the variance of the regression estimator using the generalized estimating equations approach of Zeger and Liang (1986). Paul and Zaihra (2008) conclude that the method based on an estimate of the variance of a ratio estimator performs best overall. In this paper, we extend the two new methodologies introduced in Paul and Zaihra (2008) to confidence interval construction of the risk measures RR and RED. Extensive simulations show that the method based on an estimate of the variance of a ratio estimator performs best overall for constructing confidence interval for the other two risk measures RR and RED as well. This method involves a very simple variance expression which can be implemented with a very few computer codes. Therefore, it can be considered as an easily implementable alternative for all the three measures of association.

Bayesian confidence interval for difference of the proportions in a 2×2 table with structural zero

This article studies the construction of a Bayesian confidence interval for risk difference in a 2×2 table with structural zero. The exact posterior distribution of the risk difference is derived under the Dirichlet prior distribution, and a tail-based interval is used to construct the Bayesian confidence interval. The frequentist performance of the tail-based interval is investigated and compared with the score-based interval by simulation. Our results show that the tail-based interval at Jeffreys prior performs as well as or better than the score-based confidence interval.

Estimating risk difference from relative association measures in meta-analysis can infrequently pose interpretational challenges

Objective Risk difference (RD) is often estimated from relative association measures generated by meta-analysis and a particular group's baseline risk. We describe a problematic situation in using this approach. Study Design and Setting We encountered a meta-analysis in which a confidence interval (CI) of relative risk (RR) overlapped 1.0; the point estimate favored treatment A, but when we used RR and median baseline risk to calculate a CI for RD, a greater portion of the CI favored treatment B (a result that some may find counterintuitive). We then calculated 10 different RDs from recently published meta-analyses in outcomes in which CIs of RR crossed 1.0 using three methods: estimation from RR, estimation from the odds ratio, and pooling RDs across trials. Results When RD is estimated from relative measures, the counterintuitive result occurred in 2 of 10 instances. This discordance of interpretation is because of the logarithmic transformation that makes CIs of relative measures asymmetric around their point estimates. Conclusion When RD is estimated from relative association measures that are nonsignificant and this counterintuitive situation occurs, it may be more appropriate to pool RD across studies. Pooling is particularly valid when baseline risks across studies are homogeneous.