Others

Quality Assurance and Quality Control in Longitudinal Studies

Categories
Published
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Description
Epidemiologic Reviews Copyright 1998 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. 20, No. 1 Printed in U.S.A. Quality Assurance and Quality Control in Longitudinal
Transcript
Epidemiologic Reviews Copyright 1998 by The Johns Hopkins University School of Hygiene and Public Health All rights reserved Vol. 20, No. 1 Printed in U.S.A. Quality Assurance and Quality Control in Longitudinal Studies Coralyn W. Whitney, Bonnie K. Lind, and Patricia W. Wahl INTRODUCTION Many elements of quality control for cohort studies are similar to those for other types of studies, i.e., standardization of protocol (study procedures), good communication among all study staff, clear expectations of requirements, and monitoring to ensure that requirements are met. However, cohort studies have the added dimension of longitudinal data collection, which adds issues related to drift over time, changes in equipment (including both degradation due to wear and upgrades due to new technology), and staff turnover. If the study uses central laboratories or reading centers to process data, these sites are particularly subject to quality issues related to changes over time. Cohort studies involving multiple centers have extra issues related to comparability of data collected at different locations. This presentation will provide a general overview of quality control elements common to many types of medical research studies, but primarily focuses on issues related to cohort studies. Emphasis will be placed on multicenter cohort studies, as they face additional quality control issues compared with single-center cohort studies. Much of the literature on quality assurance and quality control has arisen in the context of clinical trials, focusing on maximizing the quality of the data through standardized study-wide and local protocols, training of study personnel, and data management systems (1-9). Quality assurance and control procedures are generally described in the context of a specific study, such as the Multiple Risk Factor Intervention Trial (10-12), the Hypertension Primary Prevention Trial (3, 13), or the Optic Neuritis Treatment Trial (14, 15), with some articles providing general guidelines (4, 16-18). There is less literature in the context of nonclinical trial studies, although some study-specific approaches (8, 19-22) and general quality assurance/quality control guidelines have been Received for publication June 26, 1997, and accepted for publication May 8, From the Department of Biostatistics, School of Public Health, University of Washington, Seattle, WA. Reprint requests to Bonnie K. Lind, Department of Biostatistics, Box , University of Washington, Seattle, WA addressed (5, 23-25). The literature specifically addressing quality control issues in longitudinal studies is limited (10, 16, 26). Clinical trials often have a large number of sites, each collecting a small amount of data on a few individual patients, while cohort studies generally involve fewer sites collecting a much broader set of data. This facet of cohort studies can lead to unique quality control challenges due to the sheer bulk of the data collected (2, 10, 27, 28). Therefore, quality control literature based on clinical trials is often not relevant to cohort studies. By having fewer sites, the cohort model is more conducive to the development of long-term relationships between central quality control supervisors and site personnel, and to the supervision of individual site performance. However, it also means that the consequences of a site with quality control problems (e.g., failure to follow protocol, sloppy data handling procedures) can be much more grave for the study as a whole. Quality control is one of the most important aspects of any study, as the integrity of the conclusions drawn by a study are in large part determined by the quality of the data collected. Data of poor quality, containing a great deal of random noise, decrease the power of a study and can cause a type II error. An even worse result is the collection of data that are biased due to faulty instruments or errors in implementing the protocol, leading to an incorrect report of relations (type I error). Many aspects of data collection can impact the quality of the data, including completeness and clarity of questionnaires, the interviewer's delivery, the accuracy of mechanical instruments, and technicians' measurement techniques (16, 19, 20, 23). The validity of the study depends on the interviewers and technicians from all centers consistently applying study protocol (3, 16). Other errors can be introduced into the data after the data are collected, during transcription, at data entry, and data manipulation for analysis (1, 2, 29). Minimizing all of these potential sources of error is of paramount importance in the planning and implementation of any study. This presentation will suggest ways of minimizing these potential sources of error. There are actually two basic components to quality 71 72 Whitney et al. control: quality assurance and quality control. Quality assurance consists of those activities that take place prior to data collection while quality control consists of those activities that take place during and after data collection to identify and correct any errors or discrepancies in the data that have been collected (table 1). Each of these components will be discussed separately. QUALITY ASSURANCE Overview As stated above, quality assurance consists of those activities undertaken prior to data collection to ensure that the data are of the highest possible quality at the time they are collected. These activities include the development of the study protocol, development of the data entry and data management systems, training and certification of data collection personnel, and testing of data collection procedures. Development of the study protocol The major component of quality assurance is the development of the study protocol and the creation of manuals documenting the protocol. Aspects of protocol development that are common to most study designs are summarized below (table 1). Design of data collection instruments, including content, format, and step-by-step instructions for completion of the instrument. Procedures should be designed to ensure that the data being produced are in fact reliable, valid, on the appropriate scale, and do not reflect bias of the instrument or bias that may arise in subgroups of the population being studied (13, 14, 20, 23, 26, 30). For example, older participants tend to give responses that are socially desirable rather than strictly accurate (19). Limitations of various types of instruments should be considered during development so that potential problems can be identified and mitigated to the extent possible. For example, issues in- TABLE 1. Summary of quality assurance and quality control activities Quality assurance activities Quality control activities Development of protocol and recruitment Protocol development and design Checking for participant eligibility Questionnaires and data collection instalments Design Pilot testing Coding and editing procedures Trouble shooting Missing values Updating Data cleanup Local and central evaluation Problems with self-report data Interviewer and technician protocol Development/pilot testing Training of interviewers Training of technicians Certification procedures Recertification procedures For each interviewer and technician: Observe application of protocol Monitor certification maintenance Completeness of data Retraining when needed Regular feedback on quality control Training of new personnel Data entry packages Development/pilot testing Training of personnel Updating systems Double entry Study wide Development of local quality assurance/ quality control Monthly quality control reports Reliability/reproducibility studies Site visits Use and maintenance of study equipment Develop and pilot equipment use Training of personnel Maintenance and schedules Routine maintenance of equipment Trouble shooting Quality of equipment data Quality Assurance and Quality Control 73 herent in self-administered questionnaires have been described in the literature, such as the lack of reliability of self-reported data and the potential for questions to be misunderstood (23). Adequate testing of selfadministered questionnaires on volunteers similar to the anticipated cohort are crucial for identifying any problematic questions before the study begins. Interviewer-administered questionnaires also have potential problems; for example, participants may respond differently to interviewers of different age, gender, or ethnic background. In addition, it is vital that the protocol outlines how interviewers are to respond when participants ask for clarification of a question (i.e., is the interviewer allowed to rephrase the question in his/her own words, and if so, how does the study ascertain that interviewers are rephrasing a question in comparable ways). During planning, investigators should also consider any problems that may arise with regard to cohort members who are illiterate or do not speak English. Cohort studies typically involve many interviewers, creating the potential for differences between interviewers or centers and for differences over time (due either to drift or staff turnover). This makes the establishment of a documented, standardized interviewing protocol of paramount importance (2, 8, 10-12, 31). In addition, investigators must decide during the planning phase whether questionnaires will be allowed to change over time and how the study will ascertain that responses given over time are comparable. All aspects of the protocol must be documented in a manual of operations (2). This document should be viewed as the official study reference document for data collection staff and, as such, should contain all details of data collection procedures. Once the manual is written, it should be reviewed so that any sections that are ambiguous or subject to misinterpretation can be clarified (32). The manual should help to ascertain that data collection is consistent across field centers and over time. Many good manuals of operations exist in ongoing multicenter cohort studies, with examples usually available from funding sources (33-35). Selection of standard equipment for all measured data. All sites should begin the study with identical equipment to minimize data variability due to collection by different equipment. A protocol should be established to maintain and recalibrate equipment over time, and forms should be developed to document that these activities have occurred. Ways of identifying and replacing equipment that is worn out or failing must be agreed upon, and procedures for dealing with the development of new technologies must be discussed. Development of procedures for reviewing and updating the protocol as needed, and for communicating changes to all study personnel. Once data collection actually begins, minor adjustments to the protocol are usually needed. Oversight of this function should be assigned to a group of investigators, and a protocol should be developed for implementing and communicating such changes to all study personnel (see the subsection on Communication, below). Development of procedures for obtaining and maintaining certification to perform study procedures, and procedures for monitoring that requirements are met. An integral component of quality assurance is the training and certification of study personnel in accordance with study protocol (2, 3, 10-14, 16, 32). Procedures must be established for training and certifying data collection personnel prior to the start of the study, as well as for training and certifying new staff hired throughout the course of the study. In addition, requirements for maintaining certification must be set (both in terms of number of procedures performed and the quality of the performance), and procedures must be developed for recertification of any technicians who fail to maintain certification. Hiring individuals to serve as field center technicians is also a crucial part of quality assurance. The personalities and skills of these individuals will have a direct bearing on the final quality of study data. Some sites have found that data quality is better if technicians are hired who have little or no medical background. A person who has been performing procedures such as blood pressures in a clinical setting for many years may be more difficult to retrain to follow a standard protocol than a person who has never performed blood pressures. Communication. Communication is a key feature of successful cohort quality control systems (2), and part of the protocol development process should be devoted to setting up the structure of the communication system in the study (figure 1). Many large studies use a committee structure to ensure that tasks are accomplished and problems addressed, and in multicenter studies, the coordinating center plays the key role of facilitating communication between committees and field centers. For example, an operations committee may be formed to address questions that arise about implementation of the protocol, to ensure that the protocol is followed at all sites, and to address any problems that arise. A quality control committee may be formed to monitor the actual data quality and to address any problems that arise. For such committees to be effective, communication routes must be established. These communications often take the form of quality control reports, typically produced by the coordinating center, which are distributed to principal investigators. It is equally important that proce- 74 Whitney et al. QC Committee Operations Committee i \ ( Reading ^ \. Center J Principal Investigator i Study Coordinator Data Interviewers Technical Personnel Personnel Coordinating Center I * : Indicates direction of responsibilities FIGURE 1. Organization and communication among study personnel. dures be developed for responding to quality control reports, including identification of the person who is to respond to any problems identified in a quality control report, what steps are to be taken, and what the deadline is for response (table 2). During the development phase, the study leadership may also want to delegate responsibility for quality control at each center to a specific individual designated as the quality control supervisor, who could be an investigator, the study coordinator, or another staff member. Development of data editing, transcription, and entry procedures Collection of accurate data is only the first step. It is equally important that errors are not introduced in the process of converting the data to electronic format. For some forms it may be desirable to have the form reviewed and edited before data entry. The purpose of this review would be to ascertain that the form was complete, that skip patterns were followed, and that any data values that seemed inconsistent or looked like possible errors were checked before being entered. Once the data are ready to be converted to electronic format, several options are available. In the past, most studies have relied upon key-entry of data into the computer. Software packages exist that allow data entry to be customized to limit the range of entered data, to check for internal consistency, and to catch errors in key fields such as participant identification numbers. Many studies use double-entry verification to catch and correct any data entry errors from the original entry (16, 36-38). While this system may seem cumbersome, it is more efficient than trying to identify and correct errors at a later point in time (38). New technologies are continually being developed which offer more efficient ways of converting data to electronic format. Some studies have interviewers enter data directly into a computer without first writing the data onto a paper form (21). In this type of system, it is important that the data entry software include as many ways of checking data accuracy at entry as possible, as any errors identified later cannot be compared with a paper form for correction. Some of the newest technologies being used involve scanning paper forms directly into the computer, or faxing forms to a central data center where the fax machine serves as a scanner to convert the data to electronic format. These systems are also very efficient, but it is important that scanned files be visually reviewed for accuracy. Training and certification Once all protocols have been developed and documented, the next step is to train study personnel to implement the procedures. Training and certification activities result in standardization and can lead to reduced costs over time (16). Such training is key, as the interviewer's or technician's perceived value of the data being collected can have a direct impact on the care taken in following the protocol and the likelihood that discrepancies are introduced (27, 32). In a multicenter study, a central training session is often the best way to ensure that all personnel are trained in a standardized manner, providing the added benefit of instilling in local personnel the wider scope of the study and the importance of their work. Testing of procedures The final step in the quality assurance process is to test all procedures that have been developed. This is 2. Quality control and recruitment reports responsibilities and distribution Report Purpose Frequency/ distribution* Committee responsible Distribution From Response required Schedule ruitment and data completeness report Tabulates recruitment and study data received at coordinating center by center Weekly (F) Operations committee Principal investigators, program office, study coordinators Principal investigators As needed ed studies lity ot studies hnical data distributions tocol adherence report ruitment by demographics a cleanup Tabulate no. of failed studies Evaluation of studies by center, personnel, and any reading centers Analysis of technical procedure distributions by center and technicians Adherence to protocol: No. of technical procedures Equipment maintenance Supervisor checklist Analysis of demographics by center By center, identification of outliers and discrepancies In all data sent to coordinating center Monthly (F) Monthly (M) Monthly (M) Bimonthly (M) Monthly (F) Quarterly (M) Operations committee Quality control committee Quality control committee Operations committee Operations committee Coordinating committee/ quality control committee Principal investigators, program office, study coordinators Principal Investigators, program office, study coordinators Study coordinators Principal investigators, quality control committee, study coordinators Principal investigators, program office, study coordinators Study coordinators Principal Investigators Principal investigators Study coordinators Study coordinators Principal investigators Study coordinators Last day ot the mo distributed Last Irlday of the m distributed Last day of the mo distributed 15th of the month following the m distributed Calls/meetings ot Investigators 1 month after date distribution = sent by fax; M = sent by mall. 76 Whitney et al. often done by means of a pilot study in which the entire protocol is performed on volunteers who are similar demographically to the anticipated cohort. The pilot study should encompass all aspects of the protocol, including all interviews and measured variables, entry and transmission of data to the coordinating center, sending of samples to any reading centers or laboratories involved in the study, processing at the reading centers or laboratories, and creation and distribution of reports by the coordinating center. Some time should then be allowed for implementing refinements to the protocol between the end of the pilot study and the beginning of recruitment. QUALITY CONTROL Once data collection begins, the quality control procedures developed as part of the quality assurance process must be implemented. The goal of these quality control procedures is to identif
Search
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks