SAS Code Research Articles

Incomplete Block Designs for Comparing Test Treatments with Multiple Controls

Background: This article provides some new construction methods of partially balanced bipartite block (PBBB) designs for comparing test treatments with more than one control. Methods: Partially balanced incomplete block (PBIB) designs based on some association schemes such as group divisible association, and cyclic association are used for developing these methods of construction. Result: A catalogue of efficient PBBB designs is included for parameter values n1 (number of test treatments) ≤10, n2 (number of control treatments) ≤10 and r2 (replications of test treatments) and (replications of control treatments) ≤15 along with computed variances using SAS code.

A rank-based approach to design and analysis of pretest-posttest randomized trials, with application to COVID-19 ordinal scale data

Randomized controlled trials with a pretest-posttest design frequently yield ordered categorical outcome data. Focusing on the estimation of the win probability that a treated participant would have a better score than (or win over) a control participant, we developed methods for analysis and sample size planning for such trials. We exploited the analysis of covariance framework with the dependent variable being individual participants' win fractions at posttest and the covariate being the win fractions at pretest. The win fractions were obtained using the mid-ranks of the ordinal data. Simulation evaluation based on a recent randomized trial on COVID-19 suggests that the methods perform very well. A sample SAS code for data analysis is presented.

Parametric and nonparametric methods for confidence intervals and sample size planning for win probability in parallel-group randomized trials with Likert item and Likert scale data.

Data on the Likert scale are ubiquitous in medical research, including randomized trials. Statistical analysis of such data may be conducted using the means of raw scores or the rank information of the scores. In the context of parallel-group randomized trials, we quantify treatment effects by the probability that a subject in the treatment group has a better score than (or a win over) a subject in the control group. Asymptotic parametric and nonparametric confidence intervals for this win probability and associated sample size formulas are derived for studies with only follow-up scores, and those with both baseline and follow-up measurements. We assessed the performance of both the parametric and nonparametric approaches using simulation studies based on real studies with Likert item and Likert scale data. The simulation results demonstrate that even without baseline adjustment, the parametric methods did not perform well, in terms of bias, interval coverage percentage, balance of tail error, and assurance of achieving a pre-specified precision. In contrast, the nonparametric approach performed very well for both the unadjusted and adjusted win probability. We illustrate the methods with two examples: one using Likert item data and the other using Like scale data. We conclude that non-parametric methods are preferable for two-group randomization trials with Likert data. Illustrative SAS code for the nonparametric approach using existing procedures is provided.

Implementation of resource use measures in Medicare Advantage.

To complement the previously illustrated method to measure resource use in Medicare Advantage (MA) using Encounter data and provide technical details and SAS code to validate Encounter data and implement resource use measures in MA. 2015-2018 MA Encounter, Medicare Provider Analysis and Review (MedPAR), Healthcare Effectiveness Data and Information System (HEDIS), and Traditional Medicare (TM) claims data. Secondary data analysis. We select MA contracts with high data completeness (≤10% missing hospital stays in Encounter data and ≤±10% difference in ambulatory and emergency department visits between Encounter and HEDIS data). We randomly sample TM beneficiaries with a similar geographic distribution as MA enrollees in the selected contracts. We develop standardized prices of services using TM payments, and we measure MA resource use for inpatient, outpatient, Part D, and hospice services. We report identifiers/names of contracts with high data completeness. We provide SAS code to manage Encounter data, develop standardized prices, and measure MA resource use. Greater use and validation of Encounter data can help improve data quality. Our results can be used to inform studies using Encounter data to learn about MA performance.

A nonparametric approach to confidence intervals for concordance index and difference between correlated indices

ABSTRACT Concordance refers to the probability that subjects with high values on one variable also have high values on another variable. This index has wide application in practice, as a measure of effect size in group-comparison studies, an index of accuracy in diagnostic studies, and a discrimination index for prediction models. Herein, we provide a unified framework for statistical inference involving concordance indices for standard variables of binary, ordinal, and continuous types. In particular, we develop confidence interval procedures for a single concordance index and differences between two correlated indices. Simulation results show that procedures based on logit-transformation for a single index and Fisher’s -transformation for a difference between indices perform very well in terms of coverage and tail errors even when the sample size is as small as 30, unless the concordance is high and the standard is a binary variable for which at least 50 subjects are needed. We illustrate the procedures for a variety of standard variables with previously published data. Illustrative SAS code is provided.

A SAS code to estimate phenotypic-genotypic covariance and correlation matrices based on expected value of statistical designs to use in plant breeding.

Phenotypic-genotypic covariance and correlation have been useful in crop and animal breeding programs. In the study of diversity of natural populations and different cultivars of plants that are examined based on statistical design, estimation of genotypic-phenotypic covariance through expected value of statistical designs mean square is hard and time-consuming when the number of studied traits is high. Moreover, the lack of a program in this field and manual calculations make the estimation more complicated. Therefore, in this study, one program was developed in SAS that can be used to calculate the genotypic-phenotypic covariance matrix through the first part of the program based on the expected value of applied statistical designs mean square. Then, based on the covariance matrix computed from the previous design model, their correlation matrix was calculated using the second part of the program based on the interactive matrix language (IML) of SAS. The phenotypic-genotypic covariance matrices of the 12 studied traits of rice are calculated based on this code. This program could compute phenotypic-genotypic covariance and correlation matrices based on the expected value of any statistical designs.

Using fractional polynomials and restricted cubic splines to model non-proportional hazards or time-varying covariate effects in the Cox regression model.

The Cox proportional hazards model is used extensively in clinical and epidemiological research. A key assumption of this model is that of proportional hazards. A variable satisfies the proportional hazards assumption if the effect of that variable on the hazard function is constant over time. When the proportional hazards assumption is violated for a given variable, a common approach is to modify the model so that the regression coefficient associated with the given variable is assumed to be a linear function of time (or of log‐time), rather than being constant or fixed. However, this is an unnecessarily restrictive assumption. We describe two different methods to allow a regression coefficient, and thus the hazard ratio, in a Cox model to vary as a flexible function of time. These methods use either fractional polynomials or restricted cubic splines to model the log‐hazard ratio as a function of time. We illustrate the utility of these methods using data on 12 705 patients who presented to a hospital emergency department with a primary diagnosis of heart failure. We used a Cox model to assess the association between elevated cardiac troponin at presentation and the hazard of death after adjustment for an extensive set of covariates. SAS code for implementing the restricted cubic spline approach is provided, while an existing Stata function allows for the use of fractional polynomials.

Safety and efficacy of erythropoietin for the treatment of patients with optic neuritis (TONE): a randomised, double-blind, multicentre, placebo-controlled study

The human cytokine erythropoietin conveys neuroprotection in animal models but has shown ambiguous results in phase 2 clinical trials in patients with optic neuritis. We assessed the safety and efficacy of erythropoietin in patients with optic neuritis as a clinically isolated syndrome in a multicentre, prospective, randomised clinical trial. This randomised, placebo-controlled, double-blind phase 3 trial, conducted at 12 tertiary referral centres in Germany, included participants aged 18-50 years, within 10 days of onset of unilateral optic neuritis, with visual acuity of 0·5 or less, and without a previous diagnosis of multiple sclerosis. Participants were randomly assigned (1:1) to receive either 33 000 IU erythropoietin or placebo intravenously for 3 days as an adjunct to high-dose intravenous methylprednisolone (1000 mg per day). Block randomisation was performed by the trial statistician using an SAS code that generated randomly varying block sizes, stratified by study site and distributed using sealed envelopes. All trial participants and all study staff were masked to treatment assignment, except the trial pharmacist. The first primary outcome was atrophy of the peripapillary retinal nerve fibre layer (pRNFL), measured by optic coherence tomography (OCT) as the difference in pRNFL thickness between the affected eye at week 26 and the unaffected eye at baseline. The second primary outcome was low contrast letter acuity at week 26, measured as the 2·5% Sloan chart score of the affected eye. Analysis was performed in the full analysis set of all randomised participants for whom treatment was started and at least one follow-up OCT measurement was available. Safety was analysed in all patients who received at least one dose of the trial medication. This trial is registered at ClinicalTrials.gov, NCT01962571. 108 participants were enrolled between Nov 25, 2014, and Oct 9, 2017, of whom 55 were assigned to erythropoietin and 53 to placebo. Five patients were excluded from the primary analysis due to not receiving the allocated medication, withdrawn consent, revised diagnosis, or loss to follow-up, yielding a full analysis set of 52 patients in the erythropoietin group and 51 in the placebo group. Mean pRNFL atrophy was 15·93 μm (SD 14·91) in the erythropoietin group and 14·65 μm (15·60) in the placebo group (adjusted mean treatment difference 1·02 μm; 95% CI -5·51 to 7·55; p=0·76). Mean low contrast letter acuity scores were 49·60 (21·31) in the erythropoietin group and 49·06 (21·93) in the placebo group (adjusted mean treatment difference -4·03; -13·06 to 5·01). Adverse events occurred in 43 (81%) participants in the erythropoietin group and in 42 (81%) in the placebo group. The most common adverse event was headache, occuring in 15 (28%) patients in the erythropoietin group and 13 (25%) patients in the placebo group. Serious adverse events occurred in eight (15%) participants in the erythropoietin and in four (8%) in the placebo group. One patient (2%) in the erythropoietin group developed a venous sinus thrombosis, which was treated with anticoagulants and resolved without sequelae. Erythropoietin as an adjunct to corticosteroids conveyed neither functional nor structural neuroprotection in the visual pathways after optic neuritis. Future research could focus on modified erythropoietin administration, assess its efficacy independent of corticosteroids, and investigate whether it affects the conversion of optic neuritis to multiple sclerosis. German Federal Ministry of Education and Research (BMBF).

A note on confidence intervals for the restricted mean survival time based on transformations in small sample size.

Restricted mean survival time (RMST) is one measure now used to summarize time-to-event type data, but it has been pointed out that the distribution of differences in RMST deviates markedly from a normal distribution for controlled clinical trials with small sample sizes. Therefore, we conducted a numerical simulation of the RMST in which the one-sample survival time follows a Weibull distribution, comparing eight different confidence intervals combining two types of variance with four types of variable transformations, including no transformation. The evaluation items were the coverage probability and the above and below error probabilities for the true value. The variance types were based on Greenwood's formula and its Kaplan-Meier correction. The arcsine square root transformation, logit transformation, and complementary log-log transformation were used as the variable transformations. When the sample size was small and the event rate was low, the confidence interval of the untransformed RMST tended to have a small coverage probability and to be overestimated. Variance by Kaplan-Meier correction improved the coverage. The problems of coverage and overestimation were also improved by variable transformations, and in particular, applying the logit transformation and the complementary log-log transformation both resulted in substantial improvements. Our study suggested that it is preferable to construct the confidence intervals of RMST using the logit transformation for variances based on Greenwood's formula in small sample size trials. The SAS code to replicate the analyses is available at https://github.com/HiroyaHashimoto/SAS-Programs.

TWISTER PLOTS FOR TIME-TO-EVENT STUDIES.

Results of randomized trials and observational studies can be difficult to communicate. Results are often presented as risk or survival functions stratified by the treatment or exposure (1, 2). However, a contrast between the stratified risk functions is often of primary interest. Here we propose a “twister” plot to visualize contrasts in risk over the duration of a study. The twister plot is a −90-degree rotation of a typical contrast measure plot (e.g., Figure 3 in Cole et al. (3)), whereby the contrast measure is instead on the abscissa (x-axis) and time on the ordinate (y-axis). Pointwise confidence intervals are similarly added as a shaded region that typically widens as follow-up duration increases, giving twister plots their characteristic shape that resembles their namesake. To ease application, we provide SAS, R, and Python code on GitHub (4). In Figure 1A...

Out-of-sample earnings forecasting for OLS and Theil–Sen models relative to a na.ı.ve no-change model

PurposeThe purpose of this paper is to enlarge the exposure of the Theil–Sen (TS) methodology to the academic, analyst and practitioner communities using an earnings forecast setting. The study includes an appendix that describes the TS model in very basic terms and SAS code to assist readers in the implementation of the TS model. The study also presents an alternative approach to deflating or scaling variables.Design/methodology/approachArchival in nature using a combination of regression analysis and binomial tests.FindingsThe binomial test results support the hypothesis that the forecasting performance of the naïve no-change model is at least equal to or better than the ordinary least squares (OLS) model when earnings volatility is low. However, the results do not support the same hypothesis for the TS model nor do the results support the hypothesis that the OLS and TS models will outperform the naïve no-change model when cash flow volatility is high. Nevertheless, the study makes notable contributions to the literature, as the results indicate that the performance of the naïve model is at least as good as the OLS and TS models across 18 of the 20 binomial tests. Moreover, the results indicate that the performance of the TS model is always superior to the OLS model.Research limitations/implicationsThe results are generalizable to US firms and may not extend to non-US firms.Practical implicationsThe TS methodology is advantageous to OLS in that the results are robust to outlier observations, and there is no heteroscedasticity. Researchers will find this study to be useful given the use of a model (i.e. TS) which has to date received little attention, and the provision of the details for the mechanics of the model. A bonus for researchers is that the study includes SAS code for implementing the procedure.Social implicationsAwareness of alternative forecast methodologies could lead to improved forecasting results in certain contexts. The study also helps the financial community in general, as improved forecasting abilities are important for all capital market participants as they improve market efficiency.Originality/valueAlthough a healthy literature exists for examining out-of-sample forecasts for earnings, the literature lacks an answer for a simple question before pursuing additional analyses: Are the results any better than those from a naive no-change forecast? The current study emphasizes the idea that the naïve no-change forecast is the most elementary model possible, and the researcher must first establish the superiority of a more complex model before conducting further analyses.

On the construction of trend resistant PBIB designs

In agricultural, animal, fisheries and industrial experimentation under block design setup, systematic trend may affect the response under consideration. Although remote, these effects may still have high influence on response and hence should be incorporated in the model for proper model specification. Completely trend resistant block designs, also known as trend free block designs, are an important class of designs where the analysis can be done in usual manner as in case of normal block designs since trend effects and treatment effects are orthogonal. Literature is available on different aspects of trend resistant incomplete block designs mainly focusing on balanced incomplete block designs. This article deals with the construction of completely trend resistant Partially Balanced Incomplete Block (TR-PBIB) designs. Two methods of constructing such designs have been discussed. The information matrices for the class of designs developed have been derived and SAS code for the generation of the developed designs has also been provided.

Research tool for classifying Gulf War illness using survey responses: Lessons for writing replicable algorithms for symptom-based conditions

AimsGulf War illness (GWI), a chronic symptom-based disorder, affects up to 30% of Veterans who served in the 1990-1991 Gulf War1. Because no diagnostic test or code for GWI exists, researchers typically determine case status using self-reported symptoms and conditions according to Kansas2 and CDC3 criteria. No validated algorithm has been published and case definitions have varied slightly by study. This paper aims to standardize the application of the original CDC and Kansas case definitions by defining a framework for writing reliable code for complex case definitions, implementing this framework on a sample of 1343 Gulf War Veterans (GWVs), and validating the framework by applying the code to a sample of 41,077 GWVs. Main methodsMethods were drawn from software engineering: write pseudocode, write test cases, and write code; then test code. Code was examined for accuracy, flexibility, replicability, and reusability. Key findingsThe pseudocode promoted understanding of the planned algorithm, encouraging discussion and leading to agreement on the case definition algorithms among all team members. The completed SAS code was written for and tested in the Gulf War Era Cohort and Biorepository (GWECB)4. This code was adapted and tested in the Million Veteran Program (MVP)5. The code was documented for reproducibility and reusability. SignificanceEase of reuse suggests that this method could be used to standardize the application of other case definitions, reducing time and resources spent by each study team. Documentation, code, and test cases are available through the Department of Veterans Affairs (VA) Phenomics catalog6.

Abstract P138: Diet Quality Improvements In Response To Evening Snacks That Differ By Carbohydrate And Fat Composition: A 6-week, Randomized, Crossover Trial In Participants With Impaired Fasting Glucose

Objectives: To examine the effect of consuming one ounce of peanuts (PNUT) as an evening snack on diet quality compared to an isocaloric lower fat higher carbohydrate snack (LFHC), in individuals with impaired fasting glucose IFG. Methods: Fifty-one individuals (48% female; 42 ± 15 y; BMI 28.3 ± 5.6 kg/m 2 ; glucose 100 ± 8 mg/dL; total cholesterol 189 ± 30 mg/dL; LDL-C 121 ± 26 mg/dL; HDL-C 53 ± 14 mg/dL; triglycerides 116 ± 73 mg/dL) were enrolled in this two-period, randomized, crossover trial. In random order, subjects consumed each snack in the evening (after dinner and before bedtime) for 6 weeks (PNUT: 164 kcal, 14 g fat, 2.2 g saturated fat, 6 g carbohydrate, 7 g protein, 2.4 g fiber; LFHC: 165 kcal, 6 g fat, 2 g saturated fat, 22 g carbohydrate, 7 g protein, 3.0 g fiber) with a 4 week compliance break. Subjects were instructed not to consume other caloric foods/beverages after dinner. Participants self-reported being adherent to the protocol on 88% of study days. Dietary intake was assessed using 24-hour recalls (ASA24® Dietary Assessment Tool) conducted at the beginning and end of each diet period. The Healthy Eating Index-2015 (HEI-2015) was calculated using the NCI SAS code. Results: There was no between-condition difference in the HEI-2015 score for PNUT compared to LFHC (mean difference 3.2; 95% CI -1.1, 7.4). Individual mean component scores were significantly different following PNUT compared to LFHC (whole grains: -2.0 [95% CI -3.1, -1.0]; seafood and plant protein: 1.5 [95% CI 0.8, 2.3]; fatty acids: 2.0 [95% CI 0.8, 3.2]; and saturated fat 1.2 [95% CI 0.1, 2.4]). Following PNUT, consumption of polyunsaturated fatty acids (3 g; 95% CI 0.2, 6.6), total protein foods (2.0 oz-eq; 95% CI 0.5, 3.4) and vegetable oils (6.5 g; 95% CI 1.6, 11.5) were higher whereas whole grain (-0.9 oz-eq; 95% CI -1.2, -0.5) consumption was lower compared to LFHC. No other differences in dietary intake were observed. Conclusions: In individuals with IFG, consuming 28g of peanuts as an evening snack increased consumption of total protein foods, oils, and polyunsaturated fatty acids and reduced whole grains compared to the LFHC snack. Overall diet quality was not increased, but fatty acid and total protein food scores improved.

Confidence interval estimation for treatment effects in cluster randomization trials based on ranks.

A cluster randomization trial is one in which clusters of individuals are randomly allocated to different intervention arms. This design has become the standard for the evaluation of health care and educational strategies. To assess treatment effect, many cluster randomization trials involve outcomes that are lack meaningful units, making interpretation difficult. This difficulty may be dealt with by estimating the Mann-Whitney probability, which quantifies the probability that a typical response from one treatment arm is larger (or smaller) than a typical response from the other arm. In this work, we propose procedures for estimating this probability in cluster randomization trials. Primary emphasis is given to confidence interval estimation in trials with a small number of large clusters. The essence of the procedures is to obtain placement values based on overall ranks and arm-specific ranks prior to application of the ratio estimator, cluster-size-weighted means and mixed models for adjusting clustering effects. Nine confidence intervals were developed by applying three interval methods each based on the three variance estimators. The proposed methods can be applied to studies with binary, ordinal or continuous outcomes without making parametric assumptions. Simulation results demonstrated that the three variance estimators performed equally well, with the confidence interval procedures based on logit and inverse hyperbolic sine transformations performing better in terms of coverage and average interval width, even when the numbers of clusters are as small as 3 to 5 clusters per arm. The methods are illustrated using data from three published cluster randomization trials with SAS code provided.

Comparing Automated Cluster Detection Methods for Carbapenem-Resistant Enterobacteriaceae (CRE): Rule-Based Versus Statistical

Background: The Tennessee (TN) Department of Health (TDH) has been identifying clusters of reportable conditions using the Electronic Surveillance System for Early Notification of Community-Based Epidemics (ESSENCE), a cluster detection method using space-time scan permutation statistics based on patient ZIP code. CRE are reportable in Tennessee; isolate submission is required for carbapenemase (CP) production and resistance mechanism (eg, KPC gene) testing. The Council for Outbreak Response: Healthcare-Associated Infections (HAI) and Antimicrobial-Resistant (AR) Pathogens (CORHA) released proposed thresholds of reporting CRE to public health. Thresholds vary by healthcare facility type and regional epidemiology. The TDH HAI/AR program currently runs a daily automated SAS code using the CORHA reporting threshold to help public health identify suspect KPC clusters. We evaluated our rule-based CORHA method against 2 space-time statistic-based methods for KPC cluster detection in Tennessee. Methods: Simulations for each cluster detection method were performed using retrospective CP-CRE surveillance data for 2018. Simulations were conducted using (1) CORHA reporting thresholds by facility case count to flag clusters of 2 or more cases within 28 days, (2) ESSENCE using patient residence ZIP code and the earliest of collection date or symptom onset date as is used for other reportable conditions in Tennessee, and (3) a modified space-time statistical method using SaTScan in which reporting facility, rather than a geographic location, was used as space variable to detect within-facility clusters within 1–28 days. We compared the number and overlap of cases and clusters identified with each method. Univariate logistic regression with CORHA flagging as predictor and flagging by each ESSENCE or CORHA method as outcome variables, were used to compare cases tagged by each method pair, respectively. Results: Of 183 KPC CP-CRE cases, 54 (30.6%) were flagged as part of suspect clusters by at least 1 method. Simulations generated 16 alerts (36 cases) using CORHA, 10 clusters (25 cases) using modified SaTScan, and 10 clusters (20 cases) using standard ESSENCE protocol. Among KPC CP-CRE cases flagged by CORHA, 12 (33.3%) were also flagged by modified SaTScan and 2 (5%) by ESSENCE. A case flagged using CORHA method has 5.15 (95% CI, 2.10–12.64) times higher odds of also being flagged by the modified SaTScan method compared to cases not flagged by CORHA. Conclusions: An algorithm based on CORHA thresholds for reporting CRE to public health had strong agreement with modified SaTScan, a space-time method. We intend to explore the extension of the time interval for ESSENCE.Funding: NoneDisclosures: None

Digital Module 17: Data Visualizations: Effective Evidence‐Based Practices https://ncme.elevate.commpartners.com

AbstractIn this digital ITEMS module, Nikole Gregg and Dr. Brian Leventhal discuss strategies to ensure data visualizations achieve graphical excellence. Data visualizations are commonly used by measurement professionals to communicate results to examinees, the public, educators, and other stakeholders. To do so effectively, it is important that these visualizations communicate data efficiently and accurately. These visualizations can achieve graphical excellence when they simultaneously display data effectively, efficiently, and accurately. Unfortunately, measurement and statistical software default graphics typically fail to uphold these standards and are therefore not suitable for publication or presentation to the public. To illustrate best practices, the instructors provide an introduction to the graphical template language in SAS and show how elementary components can be used to make efficient, effective, and accurate graphics for a variety of audiences. The module contains audio‐narrated slides, embedded illustrative videos, quiz questions with diagnostic feedback, a glossary, sample SAS code, and other learning resources.

Digital Module 13: Monte Carlo Simulation Studies in Item Response Theory

AbstractIn this digital ITEMS module, Dr. Brian Leventhal and Dr. Allison Ames provide an overview of Monte Carlo simulation studies (MCSS) in item response theory (IRT). MCSS are utilized for a variety of reasons, one of the most compelling being that they can be used when analytic solutions are impractical or nonexistent because they allow researchers to specify and manipulate an array of parameter values and experimental conditions (e.g., sample size, test length, and test characteristics). Dr. Leventhal and Dr. Ames review the conceptual foundation of MCSS in IRT and walk through the processes of simulating total scores as well as item responses using the two‐parameter logistic, graded response, and bifactor models. They provide guidance for how to implement MCSS using other item response models and best practices for efficient syntax and executing an MCSS. The digital module contains sample SAS code, diagnostic quiz questions, activities, curated resources, and a glossary.

A Simple Mantel–Haenszel Type Test for Noninferiority

Randomized clinical trials designed to establish noninferiority of an experimental therapy as compared to a standard (active-control) therapy as measured by binomial proportions are being widely used. Randomization is often stratified by prognostic factors. We propose a Mantel–Haenszel (MH) type test to demonstrate noninferiority in terms of the differences in success proportions and when the noninferiority margin is not necessarily uniform in all strata. It is assumed that the efficacy of the active control is documented in at least one earlier placebo-controlled trial. Derivation of the new test originates from Wittes and Wallenstein, who have discussed asymptotic unconditional power and sample size for the MH test. We provide an easy to calculate formula for sample size. We call the new test as W-square test. It may be an alternative to the one proposed by Yanagawa et al. Yanagawa’s Neriproct suppository data are analyzed using the W-square test. A SAS code that calculates the p-value is provided.

Prospective Assessment of Chronic Lung Allograft Dysfunction (CLAD); the Clinical Trials in Organ Transplantation (CTOT) Study Experience

Diagnosis of CLAD after lung transplantation (LT) requires sustained decline in lung function and exclusion of other etiologies for the decline. The limitations of relying primarily on pulmonary function test (PFT) decline to define CLAD onset for large cohort studies are not well characterized. Using a large, prospective, multicenter cohort of LT recipients, we sought to implement a rigorous approach to identify and confirm CLAD using automated analysis of serial PFT trajectory combined with local site adjudication to rule out exclusionary conditions. The CTOT-20 cohort consists of 803 adult first LT recipients enrolled between 2015 and 2018 from 5 North American centers. PFT data transfers occurred monthly from sites to the data coordinating center, which applied an automated process using SAS code to identify potential CLAD in subjects who survived at least 90 days and had at least 5 PFTs. Potential CLAD cases, defined by sustained PFT decline per ISHLT definition, were sent to the local site to either identify alternative reasons for the decline or confirm CLAD. The site adjudication process included systematic assessment of clinical findings, bronchoscopy, microbiology, pathology, and radiographic information. 746 subjects were included in the analysis and had on average 18.9 (range 5-77) PFTs. Potential CLAD was identified a total of 379 times in 195 subjects based on PFT decline. CLAD was confirmed by the local site PI in 108/195 subjects (55.4%) and the mean time to CLAD onset was 14.8 (range 3.6-39.6) months. CLAD was confirmed even less frequently during earlier timepoints (28% in months 0-6, 43% in months 7-12, and 52% in months 13-18). The most frequent concurrent conditions excluding CLAD diagnosis were active infection (34.4%), airways stenosis (16.0%), pleural effusion/disease (14.9%), and acute rejection (14.1%). Utilizing an automated approach to identify potential CLAD supplemented by local site adjudication, we found that the majority of potential CLAD qualifying events by PFT decline had evident alternative explanations. Our results indicate that CLAD adjudication utilizing a broad spectrum of clinical information is critical. This approach was used effectively across multiple centers and provides a robust strategy that can be leveraged in future multicenter CLAD studies.