Causes Of Missing Data Research Articles

Handling missing data in a rheumatoid arthritis registry using random forest approach

Missing data in clinical epidemiological research violate the intention-to-treat principle, reduce the power of statistical analysis, and can introduce bias if the cause of missing data is related to a patient's response to treatment. Multiple imputation provides a solution to predict the values of missing data. The main objective of this study is to estimate and impute missing values in patient records. The data from the Kuwait Registry for Rheumatic Diseases was used to deal with missing values among patient records. A number of methods were implemented to deal with missing data; however, choosing the best imputation method was judged by the lowest root mean square error (RMSE). Among 1735 rheumatoid arthritis patients, we found missing values vary from 5% to 65.5% of the total observations. The results show that sequential random forest method can estimate these missing values with a high level of accuracy. The RMSE varied between 2.5 and 5.0. missForest had the lowest imputation error for both continuous and categorical variables under each missing data rate (10%, 20%, and 30%) and had the smallest prediction error difference when the models used the imputed laboratory values.

PCN223 Recommendations For The Collection Of Patient-Reported Outcomes (PRO) Data In The Post-Progression And Post-Treatment Setting Following Oncology Clinical Trials

Patient-reported outcomes (PROs) in the post-progression and post-treatment setting are important for continued understanding of the patient experience following oncology clinical trials. This research aims to summarize published literature and recommendations for the collection of PRO data post-progression and the use of this data in HTA decision-making. A targeted literature review was performed to identify publications from within the last 5 years. A search was performed on the Food and Drug Administration (FDA), European Medicines Agency (EMA) and various HTA agency websites for guidelines or recommendations. The literature review yielded 14 publications that were included in the analysis, with recommendations for protocol considerations, leveraging technology to facilitate data collection, prespecifying statistical analysis plans, and detailing causes of missing data. EMA advises that study duration should robustly support planned analysis and that PROs should be collected after progression to compare trends over time, where feasible. FDA advises that patients who fail to report due to dropout or treatment discontinuation should continue to provide PRO data, and the protocol should state how this data will be collected if withdrawal is unavoidable. The Australian Health Ministers’ Advisory Council highlights the need for further understanding of the transition between and after active treatment of cancer. The 8 HTA agencies reviewed did not provide specific recommendations for collection of this data. Continued collection of PRO data after disease progression and study treatment gives valuable insight into patients’ experiences, even as they experience changing health states. It also provides a comprehensive interpretation of disease and treatment-related impacts and is utilized by regulators and HTAs to differentiate treatments and contextualize study results, as well as patients and physicians to inform treatment selection. Current recommendations provide valuable insight, however, more guidance is needed to thoroughly inform PRO collection in this setting.

Research Attitudes Questionnaire scores predict Alzheimer's disease clinical trial dropout.

Missing data are a notable problem in Alzheimer's disease clinical trials. One cause of missing data is participant dropout. The Research Attitudes Questionnaire is a 7-item instrument that measures an individual's attitudes toward biomedical research, with higher scores indicating more favorable attitudes. The objective of this study was to describe the performance of the Research Attitudes Questionnaire over time and to examine whether Research Attitudes Questionnaire scores predict study dropout and other participant behaviors that affect trial integrity. The Research Attitudes Questionnaire was collected at baseline and weeks 26 and 52 from each member of 119 participant/study partner dyads enrolled in a Phase 2, randomized, double-blind, placebo-controlled mild-to-moderate Alzheimer's disease clinical trial. Within-subject longitudinal analyses examined change in Research Attitudes Questionnaire scores over time in each population. Logistic regression analyses that controlled for trial arm and clustering in trial sites were used to assess whether baseline Research Attitudes Questionnaire scores predicted trial completion, study medication compliance, and enrollment in optional substudies. Participants and study partners endorsed statistically similar ratings on the Research Attitudes Questionnaire that were stable over time. Participants with baseline Research Attitudes Questionnaire scores above 28.5 were 4.7 (95% confidence interval = 1.01 to 21.95) times as likely to complete the trial compared to those with lower scores. Applying the same cutoff, baseline study partner Research Attitudes Questionnaire scores were similarly able to predict study completion (odds ratio = 4.2, 95% confidence interval = 1.71 to 10.32). Using a score cutoff of 27.5, higher participant Research Attitudes Questionnaire scores predicted study medication compliance (odds ratio = 5.85, 95% confidence interval = 1.34 to 25.54). No relationship was observed between Research Attitudes Questionnaire score and participation in optional substudies. This brief instrument that measures research attitudes may identify participants at risk for behaviors that cause missing data.

Strategies for the Planning and Handling of Missing Data in Nursing Research.

Missing data are an inevitable reality in research. Nurse educators can promote proactive thinking about this topic to help avoid excessive missingness. The purpose of this article is to encourage nurses to view missing data as an accepted reality and to consider strategies for anticipating and minimizing missing data throughout the research process. The common causes of missing data and ways to minimize their occurrence are discussed, along with suggestions for adopting a statistical mindset about missing data. Rubin's framework for missingness as a random process, modern statistical methods for analyzing missing data, and recommendations for reporting also are discussed. Nurse educators and researchers should understand all aspects of missing data, including the types, occurrence, causes, potential problems, and strategies for preventing, handling, and reporting missing data. By doing so, the occurrence of missing data can be lessened, thereby minimizing various problems that can result. [J Nurs Educ. 2020;59(5):249-255.].

The Effects of Missing Data Characteristics on the Choice of Imputation Techniques

One major characteristic of data is completeness. Missing data is a significant problem in medical datasets. It leads to incorrect classification of patients and is dangerous to the health management of patients. Many factors lead to the missingness of values in databases in medical datasets. In this paper, we propose the need to examine the causes of missing data in a medical dataset to ensure that the right imputation method is used in solving the problem. The mechanism of missingness in datasets was studied to know the missing pattern of datasets and determine a suitable imputation technique to generate complete datasets. The pattern shows that the missingness of the dataset used in this study is not a monotone missing pattern. Also, single imputation techniques underestimate variance and ignore relationships among the variables; therefore, we used multiple imputations technique that runs in five iterations for the imputation of each missing value. The whole missing values in the dataset were 100% regenerated. The imputed datasets were validated using an extreme learning machine (ELM) classifier. The results show improvement in the accuracy of the imputed datasets. The work can, however, be extended to compare the accuracy of the imputed datasets with the original dataset with different classifiers like support vector machine (SVM), radial basis function (RBF), and ELMs.

Overcoming challenges to data quality in the ASPREE clinical trial

BackgroundLarge-scale studies risk generating inaccurate and missing data due to the complexity of data collection. Technology has the potential to improve data quality by providing operational support to data collectors. However, this potential is under-explored in community-based trials. The Aspirin in reducing events in the elderly (ASPREE) trial developed a data suite that was specifically designed to support data collectors: the ASPREE Web Accessible Relational Database (AWARD). This paper describes AWARD and the impact of system design on data quality.MethodsAWARD’s operational requirements, conceptual design, key challenges and design solutions for data quality are presented. Impact of design features is assessed through comparison of baseline data collected prior to implementation of key functionality (n = 1000) with data collected post implementation (n = 18,114). Overall data quality is assessed according to data category.ResultsAt baseline, implementation of user-driven functionality reduced staff error (from 0.3% to 0.01%), out-of-range data entry (from 0.14% to 0.04%) and protocol deviations (from 0.4% to 0.08%). In the longitudinal data set, which contained more than 39 million data values collected within AWARD, 96.6% of data values were entered within specified query range or found to be accurate upon querying. The remaining data were missing (3.4%). Participant non-attendance at scheduled study activity was the most common cause of missing data. Costs associated with cleaning data in ASPREE were lower than expected compared with reports from other trials.ConclusionsClinical trials undertake complex operational activity in order to collect data, but technology rarely provides sufficient support. We find the AWARD suite provides proof of principle that designing technology to support data collectors can mitigate known causes of poor data quality and produce higher-quality data. Health information technology (IT) products that support the conduct of scheduled activity in addition to traditional data entry will enhance community-based clinical trials. A standardised framework for reporting data quality would aid comparisons across clinical trials.Trial registrationInternational Standard Randomized Controlled Trial Number Register, ISRCTN83772183. Registered on 3 March 2005.

Process Variation Detection using Missing Data in a Multihospital Community Practice Anatomic Pathology Laboratory

Objectives: Barcode-driven workflows reduce patient identification errors. Missing process timestamp data frequently confound our health system’s pending lists and appear as actions left undone. Anecdotally, it was noted that missing data could be found when there is procedure noncompliance. This project was developed to determine if missing timestamp data in the histology barcode drive workflow correlated with other process variations, procedure noncompliance, or is an indicator of workflows needing focus for improvement projects. Materials and Methods: Data extracts of timestamp data from January 1, 2018, to December 15, 2018 for the major histology process steps were analyzed for missing data. Case level analysis to determine the presence or absence of expected barcoding events was performed on 1031 surgical pathology cases to determine the cause of the missing data and determine if additional data variations or procedure noncompliance events were present. The data variations were classified according to a scheme defined in the study. Results: Of 70,085, there were 7218 cases (10.3%) with missing process timestamp data. Missing histology process step data was associated with other additional data variations in case-level deep dives (P < 0.0001). Of the cases missing timestamp data in the initial review, 18.4% of the cases had no identifiable cause for the missing data (all expected events took place in the case-level deep dive). Conclusions: Operationally, valuable information can be obtained by reviewing the types and causes of missing data in the anatomic pathology laboratory information system, but only in conjunction with user input and feedback.

The (Ir)Responsibility of (Under)Estimating Missing Data.

It is practically impossible to avoid losing data in the course of an investigation, and it has been proven that the consequences can reach such magnitude that they could even invalidate the results of the study. This paper describes some of the most likely causes of missing data in research in the field of clinical psychology and the consequences they may have on statistical and substantive inferences. When it is necessary to recover the missing information, analyzing the data can become extremely complex. We summarize the experts' recommendations regarding the most powerful procedures for performing this task, the advantages each one has over the others, the elements that can or should influence our choice, and the procedures that are not a recommended option except in very exceptional cases. We conclude by offering four pieces of advice, on which all the experts agree and to which we must attend at all times in order to proceed with the greatest possible success. Finally, we show the pernicious effects produced by missing data on the statistical result and on the substantive or clinical conclusions. For this purpose we have planned to lose data in different percentage rates under two mechanisms of loss of data, MCAR and MAR in the complete data set of two very different real researchs, and we proceed to analyze the set of the available data, listwise deletion. One study is carried out using a quasi-experimental non-equivalent control group design, and another study using a experimental design completely randomized

A systematic review and development of a classification framework for factors associated with missing patient-reported outcome data.

Missing patient-reported outcome data can lead to biased results, to loss of power to detect between-treatment differences, and to research waste. Awareness of factors may help researchers reduce missing patient-reported outcome data through study design and trial processes. The aim was to construct a Classification Framework of factors associated with missing patient-reported outcome data in the context of comparative studies. The first step in this process was informed by a systematic review. Two databases (MEDLINE and CINAHL) were searched from inception to March 2015 for English articles. Inclusion criteria were (a) relevant to patient-reported outcomes, (b) discussed missing data or compliance in prospective medical studies, and (c) examined predictors or causes of missing data, including reasons identified in actual trial datasets and reported on cover sheets. Two reviewers independently screened titles and abstracts. Discrepancies were discussed with the research team prior to finalizing the list of eligible papers. In completing the systematic review, four particular challenges to synthesizing the extracted information were identified. To address these challenges, operational principles were established by consensus to guide the development of the Classification Framework. A total of 6027 records were screened. In all, 100 papers were eligible and included in the review. Of these, 57% focused on cancer, 23% did not specify disease, and 20% reported for patients with a variety of non-cancer conditions. In total, 40% of the papers offered a descriptive analysis of possible factors associated with missing data, but some papers used other methods. In total, 663 excerpts of text (units), each describing a factor associated with missing patient-reported outcome data, were extracted verbatim. Redundant units were identified and sequestered. Similar units were grouped, and an iterative process of consensus among the investigators was used to reduce these units to a list of factors that met the guiding principles. The list was organized on a framework, using an iterative consensus-based process. The resultant Classification Framework is a summary of the factors associated with missing patient-reported outcome data described in the literature. It consists of 5 components (instrument, participant, centre, staff, and study) and 46 categories, each with one or more sub-categories or examples. A systematic review of the literature revealed 46 unique categories of factors associated with missing patient-reported outcome data, organized into 5 main component groups. The Classification Framework may assist researchers to improve the design of new randomized clinical trials and to implement procedures to reduce missing patient-reported outcome data. Further research using the Classification Framework to inform quantitative analyses of missing patient-reported outcome data in existing clinical trials and to inform qualitative inquiry of research staff is planned.

Maintenance Keeps Radars Running

Abstract Weather radars provide us with colorful images of storms, their development, and their movement, but from time to time the radars fail and we are left without data. To minimize these disruptions, owners of weather radars carry out preventive maintenance. The European radar project Operational Programme for the Exchange of Weather Radar Information (OPERA) conducted a survey among technicians from 21 countries on their experiences of maintenance. Regular maintenance frequency varies widely from as frequent as weekly to as infrequent as 6 months. Results show that the primary causes of missing data are not the failure of radar components and software or lack of maintenance but rather issues with the electricity supplies or telecommunications. Where issues are with the radars themselves, they are most commonly with the transmitter or the antenna controllers. Faults can be repaired quickly, but, if certain parts are required or the site is very remote, a radar can be out of service for weeks or even months. Failures of electricity or communications may also lead to lengthy periods of unavailability. As an example there is a story from Norway where wintertime thunderstorms severely damaged a radar at a very remote location. Annual operative costs of a radar are typically on the order of 5%–10% of the radar purchase price. During the lifetime of a system (typically 10–20 years) the operator can hence pay as much for the running costs as for the hardware purchase. It is extremely important to take infrastructure, maintenance, and monitoring into account when purchasing a new radar.

Identifying research priorities for effective retention strategies in clinical trials

BackgroundThe failure to retain patients or collect primary-outcome data is a common challenge for trials and reduces the statistical power and potentially introduces bias into the analysis. Identifying strategies to minimise missing data was the second highest methodological research priority in a Delphi survey of the Directors of UK Clinical Trial Units (CTUs) and is important to minimise waste in research. Our aim was to assess the current retention practices within the UK and priorities for future research to evaluate the effectiveness of strategies to reduce attrition.MethodsSeventy-five chief investigators of NIHR Health Technology Assessment (HTA)-funded trials starting between 2009 and 2012 were surveyed to elicit their awareness about causes of missing data within their trial and recommended practices for improving retention. Forty-seven CTUs registered within the UKCRC network were surveyed separately to identify approaches and strategies being used to mitigate missing data across trials.Responses from the current practice surveys were used to inform a subsequent two-round Delphi survey with registered CTUs. A consensus list of retention research strategies was produced and ranked by priority.ResultsFifty out of seventy-five (67%) chief investigators and 33/47 (70%) registered CTUs completed the current practice surveys. Seventy-eight percent of trialists were aware of retention challenges and implemented strategies at trial design. Patient-initiated withdrawal was the most common cause of missing data. Registered CTUs routinely used newsletters, timeline of participant visits, and telephone reminders to mitigate missing data. Whilst 36 out of 59 strategies presented had been formally or informally evaluated, some frequently used strategies, such as site initiation training, have had no research to inform practice.Thirty-five registered CTUs (74%) participated in the Delphi survey. Research into the effectiveness of site initiation training, frequency of patient contact during a trial, the use of routinely collected data, the frequency and timing of reminders, triggered site training and the time needed to complete questionnaires was deemed critical. Research into the effectiveness of Christmas cards for site staff was not of critical importance.ConclusionThe surveys of current practices demonstrates that a variety of strategies are being used to mitigate missing data but with little evidence to support their use. Six retention strategies were deemed critically important within the Delphi survey and should be a primary focus of future retention research.

Feasibility of Patient Reporting of Symptomatic Adverse Events via the Patient-Reported Outcomes Version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE) in a Chemoradiotherapy Cooperative Group Multicenter Clinical Trial

To assess the feasibility of measuring symptomatic adverse events (AEs) in a multicenter clinical trial using the National Cancer Institute's Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE). Patients enrolled in NRG Oncology's RTOG 1012 (Prophylactic Manuka Honey for Reduction of Chemoradiation Induced Esophagitis-Related Pain during Treatment of Lung Cancer) were asked to self-report 53 PRO-CTCAE items representing 30 symptomatic AEs at 6 time points (baseline; weekly ×4 during treatment; 12weeks after treatment). Reporting was conducted via wireless tablet computers in clinic waiting areas. Compliance was defined as the proportion of visits when an expected PRO-CTCAE assessment was completed. Among 226 study sites participating in RTOG 1012, 100% completed 35-minute PRO-CTCAE training for clinical research associates (CRAs); 80 sites enrolled patients, of which 34 (43%) required tablet computers to be provided. All 152 patients in RTOG 1012 agreed to self-report using the PRO-CTCAE (median age 66years; 47% female; 84% white). Median time for CRAs to learn the system was 60minutes (range, 30-240minutes), and median time for CRAs to teach a patient to self-report was 10minutes (range, 2-60minutes). Compliance was high, particularly during active treatment, when patients self-reported at 86% of expected time points, although compliance was lower after treatment (72%). Common reasons for noncompliance were institutional errors, such as forgetting to provide computers to participants; patients missing clinic visits; Internet connectivity; and patients feeling "too sick." Most patients enrolled in a multicenter chemoradiotherapy trial were willing and able to self-report symptomatic AEs at visits using tablet computers. Minimal effort was required by local site staff to support this system. The observed causes of missing data may be obviated by allowing patients to self-report electronically between visits, and by using central compliance monitoring. These approaches are being incorporated into ongoing studies.

MISSING DATA IMPUTATION IN THE ELECTRONIC HEALTH RECORD USING DEEPLY LEARNED AUTOENCODERS.

Electronic health records (EHRs) have become a vital source of patient outcome data but the widespread prevalence of missing data presents a major challenge. Different causes of missing data in the EHR data may introduce unintentional bias. Here, we compare the effectiveness of popular multiple imputation strategies with a deeply learned autoencoder using the Pooled Resource Open-Access ALS Clinical Trials Database (PRO-ACT). To evaluate performance, we examined imputation accuracy for known values simulated to be either missing completely at random or missing not at random. We also compared ALS disease progression prediction across different imputation models. Autoencoders showed strong performance for imputation accuracy and contributed to the strongest disease progression predictor. Finally, we show that despite clinical heterogeneity, ALS disease progression appears homogenous with time from onset being the most important predictor.

Pragmatic approaches to mitigating missing data and research priorities to assess the effectiveness of interventions

Background Identifying interventions to minimise missing data was the third highest research priority in a Delphi survey of the Directors of UK Clinical Trial Units (CTUs). However, a Cochrane Methodology Review of nested randomised studies of missing data interventions shows a substantial evidence gap, with all but one of the eligible studies targeting questionnaire response rates. Research is needed to identify strategies that effectively address the broader causes of missing data, including minimising withdrawals, ensuring clinical staff capture vital outcome measurements and improving patient attendance at clinic visits.

Multiple Imputation: A Flexible Tool for Handling Missing Data.

In this issue of JAMA, Asch et al1 report results of a cluster-randomized clinical trial designed to evaluate the effects of physician financial incentives, patient incentives, or shared physician and patient incentives on low density lipoprotein cholesterol (LDL-C) levels among patients with high cardiovascular risk. Because 1 or more follow-up LDL-C measurements were missing for approximately 7% of participants, Asch et al used multiple imputation (MI) to analyze their data and concluded that shared financial incentives for physicians and patients, but not incentives to physicians or patients alone, resulted in the patients having lower LDL-C levels. Imputation is the process of replacing missing data with 1 or more specific values, to allow statistical analysis that includes all participants and not just those who do not have any missing data. Missing data are common in research. In a previous JAMA Guide to Statistics and Methods, Newgard and Lewis2 reviewed the causes of missing data. These are divided into 3 classes: 1) missing completely at random, the most restrictive assumption, indicating that whether a data point is missing is completely unrelated to observed and unobserved data; 2) missing at random, a more realistic assumption than missing completely at random, indicating whether a data point is missing can be explained by the observed data; or 3) missing not at random, meaning that the missingness is dependent on the unobserved values. Common statistical methods used for handling missing values were reviewed.2 When missing data occur, it is important to not exclude cases with missing information (analyses after such exclusion are known as complete case analyses). Single-value imputation methods are those that estimate what each missing value might have been and replace it with a single value in the data set. Single-value imputation methods include mean imputation, last observation carried forward, and random imputation. These approaches can yield biased results and are suboptimal. Multiple imputation better handles missing data by estimating and replacing missing values many times.

Weighting and Imputation for Missing Data in a Cost and Earnings Fishery Survey

Surveys of fishery participants are often voluntary and, as a result, commonly have missing data associated with them. The two primary causes of missing data that generate concern are unit non-response and item non-response. Unit non-response occurs when a potential respondent does not complete and return a survey, resulting in a missing respondent. Item non-response occurs in returned surveys when an individual question is unanswered. Both may lead to issues with extrapolating results to the population. We explain how to adjust data to estimate population parameters from surveys using two of the principal approaches available for addressing missing data, weighting and data imputation, and illustrate the effects they have on estimates of costs and earnings in the Alaska charter boat sector using data from a recent survey. The results suggest that ignoring missing data will lead to markedly different results than those estimated when controlling for the missing data.

Missing not at random models for masked clinical trials with dropouts.

Missing data are an unavoidable problem in clinical trials. Most existing missing data approaches assume the missing data are missing at random. However, the missing at random assumption is often questionable when the real causes of missing data are not well known and cannot be tested from observed data. We propose a specific missing not at random assumption, which we call masked missing not at random, which may be more plausible than missing at random for masked clinical trials. We formulate models for categorical and continuous outcomes under this assumption. Simulations are conducted to examine the finite sample performance of our methods and compare them with other methods. R code for the proposed methods is provided in supplementary materials. Simulation studies confirm that maximum likelihood methods assuming masked missing not at random outperform complete case analysis and maximum likelihood assuming missing at random when masked missing not at random is true. For the particular missing at random model where both of missing at random and masked missing not at random are satisfied, theory suggests that maximum likelihood assuming missing at random is at least as efficient as maximum likelihood assuming masked missing not at random. However, maximum likelihood assuming masked missing not at random is nearly as efficient as maximum likelihood assuming missing at random in our simulated settings. We also applied our methods to the TRial Of Preventing HYpertension study. The missing at random estimated treatment effect and its 95% confidence interval are robust to deviations from missing at random of the form implied by masked missing not at random. Methods based on the masked missing not at random assumption are useful for masked clinical trials, either in their own right or to provide a form of sensitivity analysis for deviations from missing at random. Missing at random analysis might be favored on grounds of efficiency if the estimates based on masked missing not at random and missing at random are similar, but if the estimates are substantially different, the masked missing not at random estimates might be preferred because the mechanism is more plausible.

How do MIS researchers handle missing data in survey-based research: A content analysis approach

Missing data is a common occurrence in survey-based research studies. However, the issue of missing data in Management Information Systems (MIS) literature has been overlooked, unlike the case in other disciplines such as Psychology, Marketing, Statistics, and Operations Management. The aim of this paper is to narrow this gap in the MIS field by investigating how MIS researchers address issues of missing data. This paper briefly outlines the causes of missing data in survey-based research as well as the common remedial techniques available to researchers. The paper also reviews how the common statistical software programs namely PASW (SPSS), SAS, LISREL, AMOS, EQS, and PLS handle missing data. It summarizes the common missing data remedial techniques and procedures and outlines how the presence of missing data affect sample size, statistical power, parameter estimates, ability to cope with different missing data patterns, and ease of implementation. Following that is a review of 749 survey-based research articles published between 1990 and 2010 in nine mainstream MIS Journals. The results reveal that researchers rarely report, explicitly, the presence or treatment of missing data and that when they do – they tend to use the least accurate techniques of listwise and pairwise deletion. The research concludes with recommendations that include a call for editorial policies that encourage the reporting of missing data, the reporting of the chosen missing data treatment techniques as well as the justifications for the techniques adopted by the researchers. The authors assert, based on the research, that following these recommendations will affect the rigor and quality of MIS survey-based research.

Missing Data in Confirmatory Clinical Trials

Missing data for key efficacy and safety endpoints in clinical trials have the potential to undermine the scientific integrity of the study and prevent definitive conclusions regarding the safety and efficacy of an experimental product. Much of the missing data is the result of poor protocol design and a lack of agreement in the scientific community regarding the collection of study data after treatment discontinuation instead of an inability to collect the data. Rather than dealing with the fundamental causes of missing data, the statistical community has traditionally attempted to explicitly impute the missing data based upon observed data or more recently through the use of statistical models that implicitly impute the missing data. In this article, the causes for missing data are described and a number of approaches to maintain the integrity of the studies are described.

Missing Patient-Reported Outcome Data in an Adjuvant Lung Cancer Study

Missing Patient-Reported Outcome Data in an Adjuvant Lung Cancer Study