The Role of Interim Analysis as a Quality Assurance Function in Practice-Based Research Network Clinical Studies
Frederick A. Curro, DMD, PhD; Ronald G. Craig, DMD, PhD; Donald Vena; Van P. Thompson, DDS, PhD
In this issue of the Compendium, the authors present “Postoperative Hypersensitivity in Class I Resin-Bonded Composite Restorations in General Practice: Interim Results.” This clinical study is being conducted by the dental practice-based research network (PBRN) termed Practitioners Engaged in Applied Research and Learning (PEARL) Network, which is a partnership encompassing the National Institute of Dental and Craniofacial Research (NIDCR) as the funding source, The EMMES Corporation as the data coordinating center, and New York University College of Dentistry as the administrative core.1
The PBRN concept challenges the traditional paradigm of clinical studies. The Vioxx® (Merck & Co., Inc., Whitehouse Station, NJ) issue, which caused Merck to withdraw the drug in 2004, raised many concerns regarding safety for a recently approved drug. The number of patients in the Phase III pivotal studies totaled more than 8000, yet the full impact of Vioxx on cardiovascular health had been found insignificant. The issues that emerged once the drug was on the market initiated a complete review of the Food and Drug Administration’s (FDA) drug safety program, leading to consideration of alternatives for conducting clinical studies.
One such pathway involves research performed by clinicians in their offices in an effort to broaden the clinical population and to be more inclusive of various health conditions to better expose potential drug safety issues. The traditional drug development paradigm is sequential, consisting of animal studies, Phase I and Phase II research in humans, and ultimately randomized controlled trials (RCTs). The word trial is used instead of study when a clinical study is referenced in the sequence of a drug development program and the study involves humans. The terms investigator and subject (or participant) are reserved for RCTs while practitioner and patient are used for PBRN-conducted studies. One major criticism of RCTs is the fact that they may be too controlled and issues of drug safety may be overlooked during the course of the trials. RCTs are usually conducted for the Phase III drug development program and controlled through the use of experienced investigators in their respective fields of interest. Subjects are recruited based on rigid inclusion/exclusion criteria to minimize the impact of factors other than the treatment or procedure being studied and to limit the serious adverse reactions and/or events. RCTs are conducted using as few investigators as possible to standardize the treatment or procedure and to keep subject-response variability as low as possible. Generally, each investigator of a RCT recruits a rather large number of participants.
Studies conducted under the PBRN concept require a different strategy because the research uses the opposite paradigm—a large number of investigators called practitioners with each recruiting a small number of subjects, or patients. Generally, the studies are not investigational in the sense of using new drug entities or experimental devices; rather, they compare best practice and/or effectiveness that include the patients’ input in what the FDA terms patient-reported outcomes.2 Mostly, such research can be termed standard of care studies, and the practitioners generally are naïve to the principles of conducting research. By using a large number of naïve practitioners who recruit patients from their own practices, the data collected can be viewed as more reflective of real life but generated from a less controlled environment. PBRN study results are more reflective of the population at large and probably represent a wider potential spectrum of adverse reactions and events than those reported in RCTs. PBRN results are gaining favor, given the number of drugs recently removed from the market, because of an increased side effect profile occurring in the general population and not captured in the pivotal clinical trials.
The question of how to control the quality of the data generated by a PBRN study using a large number of relatively naïve practitioners can be addressed in several ways, some operational but not practical and others less intrusively. One way is extensive oversight by clinical research associates to monitor the study’s progress and to ensure compliance. However, this method becomes intrusive for the practitioners and their office staff and may eventually jeopardize the study’s success. The goal of the PBRN is not to make researchers out of practitioners but to conduct quality research in the practice setting where treatment is ultimately being delivered. The hallmark of a PBRN study is that it is designed to be incorporated into the logistical operational flow of the office. RCTs are conducted by experienced investigators who are reimbursed for much more than for their time involved. The added cost of such oversight can be a limiting factor and, in addition, changes the tenets of a PBRN study: “What is normally performed in your office?” or “What is the standard of care for your practice?”
Another minimally intrusive alternative is to address the quality assurance of an ongoing study by conducting interim assessments of the data. This method maintains the principles of the PBRN concept while creating an oversight procedure that does not burden the practitioner or office staff. Publishing the interim assessment creates a transparency in the conduct of the PBRN concept and adds a level of compliance to the management of the study. Interim assessment assures the administrative core of the PBRN that the protocol is being properly followed as measured by the number of queries and review of the case report forms. It prevents the study from “drifting” from its intent and becoming too “office-directed”—in other words, having office procedure supersede the procedure as stated in the protocol. It ensures that the case report forms are completed properly and patient consent has been obtained correctly. In addition, publishing abstracts or a report of an interim assessment helps maintain the interest and enthusiasm of the participating practitioners and gives them a sense of direction and security regarding their own study procedures. These issues all confront experienced investigators, and this is well documented in the drug development process. Industry remuneration can go a long way to sustain clinician interest. While it is inevitable that a PBRN consisting of naïve practitioners has a certain percentage of practitioner turnover, establishing quality control without compromising the study’s intent or data integrity can be resolved by performing an interim assessment.
The reported study in this issue of the Compendium was PEARL’s first supported clinical study and introduced the practitioners to a number of clinical research steps, including consent of the patient; review of the inclusion/exclusion criteria; measurement of the length, width, and depth of a preparation; recording of materials used to line and fill the preparation; completion of case report forms; and electronic transmittal of the data to the coordinating center for analysis. Most importantly, this study included patient follow-up forms to be sent posttreatment at weeks 1, 4, and 13. The interim assessment proved to be highly insightful for PEARL’s administrative core in regard to assurance of practitioner compliance. The interim assessment can, in itself, become a study to evaluate long-term office and practitioner compliance and especially patient-reported outcomes. The number of interim assessments becomes dependent on the study size and duration.
The study reports the procedures performed by the practitioner, using resin-based composites, as well as the patient-reported outcomes at 1 week postoperatively. The information reported from this study can be used as a basis for further assessment and optimization of treatment plans.
As noted above, PBRN effectiveness research looks at outcomes in many patients in practice-based settings. Typically, studies are powered to investigate outcomes based on re-ports in the literature. For the postoperative-hypersensitivity study, this was 610 patients. However, in review of the data at 250 patients and then at 400-plus patients reported here, the authors have noted at both end points two factors. First, upon opening the enamel of these early lesions only slightly, more than one third were ranked as caries active. Second, patients reported a much higher level of hypersensitivity than anticipated, both before and after res-toration. The authors believe that these interim findings provide important information to the dentist for reconsideration of aggressive treatment of these lesions. Also, because significant numbers of patients experience postoperative hypersensitivity who did not report any sensitivity prior to treatment, review of treatment techniques and methods is warranted. Given the importance of these interim findings, the authors believe that the results should be reported as soon as possible to the profession.
References
1. Curro FA, Craig RG, Thompson VP. Practice-based research networks and their impact on dentistry: creating a pathway for change in the profession Compend Contin Educ Dent. 2009;30(4):186-187.
2. US Department of Health and Human Services, Food and Drug Administration. Guidance for Industry: Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims. Food and Drug Administration. Rockville, MD: 2006.
About the Authors
Frederick A. Curro, DMD, PhD
PEARL Executive Management Team
Director
Regulatory Affairs and Clinical Investigations
Bluestone Center for Clinical Research
New York University College of Dentistry
New York, New York
Ronald G. Craig, DMD, PhD
PEARL Executive Management Team
Associate Professor
Department of Basic Sciences and Craniofacial Biology
Department of Periodontology and Implant Dentistry
New York University College of Dentistry
New York City, New York
Donald Vena
PEARL
Coordinating Center Principal
The EMMES Corporation
Rockville, Maryland
Van P. Thompson, DDS, PhD
PEARL Executive Management Team
Chair of Biomaterials and Biomimetics
New York University College of Dentistry
New York, New York
