Early Identification of Poorly Performing Joint Replacements

点击翻译按钮获取中文摘要

Joint replacement surgery is the most common inpatient surgery in the Veterans Health Administration (VHA). 14,349 primary joint replacement cases were performed in the VHA in 2015. About 5% are reoperated by 10 years for failure or infection. Poor implants have a much higher reoperation rate (e.g. 17-25% by 5 years for the DePuy ASR XL). Failed or unsatisfactory joint replacements are costly and result in personal and financial hardship for veteran patients. In addition, failed joint replacements cost the health care system about $38,000 per episode. As implant designs and techniques evolve, some changes can have a negative effect on implant survival and harm patients. A prime example of this is metal on metal hip arthroplasty, which recently resulted in recall of several commonly used implants. To limit harm to patients and costs for the health care system, we must improve our ability to monitor implant performance and identify poorly performing implants earlier than is now possible. Medical records are a potentially rich source of information that may allow early identification of poorly performing implants even before they end up having to be revised, and potentially sparing new patients from receiving them. The VHA is the nation's largest integrated health care system and was among the earliest adopters of the electronic health record (EHR). VHA EHRs may be used to create a VA joint implant registry. This could inform future decisions on implant selection across the VA health care system to improve veteran patient outcomes and minimize reoperations or complications. In addition, a VA joint replacement registry could act to support a national implant surveillance system by computationally extracting and aggregating individual prosthetic data for use by the FDA. We propose a pilot project to extract and evaluate joint replacement prosthesis implantation and explantation information from the VA Corporate Data Warehouse (CDW). The specific aims of the proposed pilot work are: Aim 1: Determine trends in national joint replacement implant use and regional variation in the VA from 2000 to 2016. To establish a foundation for our overall goal of developing a post market surveillance system for joint replacement, it will be essential to retrieve implant-specific information from the CDW across the VHA over time, including Operative Date, Station, CPT code, Prosthesis Vendor, Prosthesis Model, Prosthesis Size, and Prosthesis Name. After evaluating the completeness of these records across the VHA over time, we will use these records to characterize implantation practices for each station and each calendar year. This will support further study by quantifying trends in national joint replacement implant use and provide insight into regional variation in implant selection across the VA system. Aim 2: Build a level 1 registry of joint replacement implants in the VA system. A level 1 registry is made up of patient demographics, hospital, surgeon, and procedure-related data including the laterality of surgery, date of surgery and implant utilized. Not all of this information is available in structured records. For example since laterality of the operation is not directly found in a structured field, we will need to analyze unstructured fields to determine the fate of the index operation. Part (a) of this aim will focus on extracting laterality information for both index surgery and any subsequent reoperations from unstructured fields. This will tell us how long the joint prosthesis lasted before it needed to be reoperated on. Part (b) of the aim will require a more challenging analysis of unstructured data to determine which component of the joint replacement was problematic and removed. By completing this aim, we will generate reports on the individual component survivorship free of reoperation or removal for each of the implants used across the VA health care system.