Development of a coordinated registry network for pelvic organ prolapse technologies

Overview

The POP working group consisted of 21 expert members, all POP specialists from Female Pelvic Medicine and Reconstructive Surgery, including both gynecology and urology backgrounds and technology backgrounds. The full list of working group members is reported in online supplemental file 1. The working group co-chairs reduced an initial list of >300 potential data elements to 120 data elements that were included in the Delphi process. Participation rates in the first and second round of the Delphi survey were 95.2% and 71.4%, respectively. Completion of the Delphi surveys resulted in a consensus among the responders on the selection of 90 data elements identified as relevant to POP surgical devices (online supplemental file 1). The data elements were grouped in the following categories: (1) medical history; (2) surgical history; (3) examination; (4) procedure; (5) discharge; (6) short-term follow-up (0–30 days); (7) long-term follow-up (>90 days); (8) device factors; (9) surgery factors and (10) surgeon factors. Patient demographic variables (age, race, etc) were not included in the Delphi selection process as a standard, harmonized set of demographic variables were selected based on work already conducted by a multistakeholder project sponsored by the Pew Charitable Trusts.30 Subjective measures were not included as a working group has been tasked with a means to collect patient reported outcomes to be included in the registry. The level of consensus for both the final data elements and the dropped data elements are reported in online supplemental files 2 and 3.

Medical and surgical history

For the patient’s medical history, there was consensus among participants to capture overall parity, vaginal births, cesarean births, smoking status, menopause, sexual activity and pain, chronic constipation, estrogen therapy, and vaginal bulge symptoms. Lower urinary tract symptoms included mixed, stress, and urgency urinary incontinence. These were determined to be particularly important by the group as they represent risks for worsening POP or potential sequalae of POP repair. The comorbidities are best captured using a combination of the American Society of Anesthesiologists (ASA) physical status classification31 and including specific comorbidities such as diabetes mellitus. Prior surgical history included prior hysterectomy, prior mesh use, prior anti-incontinence surgery, prior prolapse surgery, and prior abdominal surgery. For physical examination parameters, body mass index, Pelvic Organ Prolapse Quantification (POP-Q) stage, and compartment with the greatest anatomic prolapse were included. This would allow for comparisons among surgical repairs as higher degrees of prolapse are more likely to result in recurrence. The hymen is an important ‘cut-off point’, as women with prolapse beyond the hymen have more pelvic floor symptoms and are more likely to report a vaginal bulge than women with prolapse above the hymen.32–35 POP stages can also be evaluated using the Baden-Walker, a clinical system that grades the degree of prolapse from 0 to 4 in three different pelvic compartments. However, this was not chosen for inclusion. All members agreed that the POP-Q assessment is the most accurate and reproducible way to measure prolapse and has the highest intersurgeon reliability.36

Procedure and discharge

Surgery date, total operative time, ASA status, concomitant hysterectomy, concomitant anti-incontinence procedure, and mesh use were determined to be core data elements related to procedure. Furthermore, the type of vaginal or abdominal apical vault suspension used, if anterior, enterocele, or posterior repair was performed, and if an obliterative procedure was performed were included. This is of particular importance since both transvaginal and transabdominal routes are commonly used and each has its own risk profiles. The complications include bleeding requiring transfusion, ureteral injury, urethrotomy, vascular injury, visceral injury, mesh kit trocar injury, other operative complication/injury, aborted procedure, conversion to laparotomy, device malfunction, and death. The most severe complication would be classified using the Clavien-Dindo classification system37 to grade the level of severity of the adverse event and to capture the need for subsequent therapy. Given that a number of complications such as suture exposure or erosion may be managed in either the office, operating room, or emergency room, the Clavien-Dindo system allows for discrimination between interventions performed with and without general anesthesia. It is important to also capture re-operation during index hospitalization, discharge date, and discharge disposition.

Short-term and long-term follow-up

For short-term follow-up (<30 days), there was consensus to capture early complications which include cardiovascular complications, pulmonary complications, systemic infections, bleeding, organ injury, suture exposure or erosion in the vagina or viscera, foreign body or death. For short-term follow-up (31–90 days), complications to be collected include vaginal scarring, shortening, suture exposure, erosion, mesh exposure or erosion, pelvic pain, dyspareunia, fistula, bowel injury, thrombolic event, cardiac event, pulmonary event, and neurovascular event. For long-term complications (>90 days), these would include the same elements as 31–90 days complications with the addition of symptomatic or anatomic recurrence. If there is a recurrence, it is appropriate to grade using the POP-Q system as well as reporting the compartment with the greatest anatomic prolapse. It is important to capture readmissions within 30 and 90 days. For all short-term and long-term complications, the most severe complication would be graded using the Clavien-Dindo scale.

Device factors

To accurately capture the device used for prolapse repair, the device identifier of the unique device identifier (UDI-DI)38 would be collected for any implant or suture used in a procedure. One or more of the production identifiers of the UDI (UDI-PI) would also be included if they appeared on the device label. The parts of the UDI-PI are: lot, serial number, expiration date, manufacturing date, distinct identification code. In addition, the UDI-DI would be used to pull data from AccessGUDID to auto-populate the company name, brand, clinically relevant size, device type (eg, absorbable, permanent), and other relevant device identification characteristics with master data across all registries.39

Surgery and surgeon factors

Surgery level data elements include practice type, center or hospital identifiers, hospital volume, and whether there was trainee involvement. Surgeons would be classified based on the National Provider Identifier, age, level of training, specialty, board certification, subspecialty certification, and surgeon volume. The panel agreed that surgical errors and a learning curve likely contribute to the rate of complications.40 The learning curve in surgery is a well-defined phenomenon in many other clinical areas and most of these studied operations are fairly large and complex.41 There are very few, if any, high-quality studies looking at the role that the learning curve plays in POP surgery. In the realm of stress incontinence, which has not been subject of FDA action, Welk et al,40 measured the incidence of mesh removal or revision after sling mesh procedures in 59 887 women over a 10-year period. The authors found that 1307 women (2.2%) underwent mesh removal or revision a median of 0.94 years after receiving a mesh implant for SUI. Patients of high-volume surgeons (75th percentile of yearly mesh-based procedures) had a significantly lower risk for experiencing reintervention (removal or revision) with no difference found among specialties (urologist vs gynecologists). These findings support the group’s decision to include surgeon and facility level data for POP and that a learning curve exists.

Informatics work

Achieving consensus on the core minimum dataset for POP was an important first step in our goal of creating a CRN for numerous women’s health conditions (WHT-CRN). Core minimum datasets were concurrently being developed for SUI, uterine fibroids, sterilization, and long-acting and reversible contraception. In order to create a CRN capable of evaluating medical devices used for all of these conditions, it was imperative to harmonize common data elements among all of the clinical areas to ensure interoperability of datasets stemming from future registries. As such, the informatics team compared, identified, and aggregated data elements under common concepts that occurred in at least two of the clinical areas for harmonization. The common concepts were intended to uncover potential gaps. Concurrently, a search was done for the common concepts using the National Institutes of Health (NIH) Common Data Elements (CDE) Repository, a platform that enables linkage of data elements to existing standards and terminologies and acted as input for initial modeling of the unique data elements. Codes were drawn from standard clinical vocabularies and other resources, such as the Value Set Authority Center, which is a repository of codes and terms from LOINC, SNOMED CT, International Classification of Diseases (ICD)-9, and ICD-10, among others and an authoring tool for public value sets. We also gathered potentially relevant Current Procedural Terminology (CPT), Healthcare Common Procedure Coding System (HCPCS), ICD-9, and ICD-10 codes for each clinical condition through online searches and referencing the corresponding codebooks. In addition, the potential UDI-DI for each of the devices used in a clinical area were validated against AccessGUDID. The data elements were modeled using forms associated with each corresponding clinical group in the NIH CDE Repository. Each form contained the full set of harmonized set of data elements and associated permissible values linked to standardized codes and the data elements required in appropriate Health Level Seven International (HL7) profiles.

The WHT-CRN Implementation Guide (IG) builds on interoperable data exchange standards such as the HL7 profiles that can be used to define WHT-CRN data infrastructure. The WHT-CRN IG focuses on capturing data related to women’s health devices and making that data available for exchange to both providers and authorized researchers. The informatics team worked with one existing women’s health registry to pilot the WHT-CRN Fast Healthcare Interoperability Resources (FHIR^®) IG and its underlying standards and datasets in a test environment. The input from that testing environment will be used to inform work with other WHT-CRN test and production environments (eg, clinical or provider settings) and/or manufacturing setting. Please see figure 1A,B for illustrations of how the WHT-CRN data will be collected and accessed by organizations. An additional outcome of testing will be to evaluate the ability to capture UDI from scanning the device label, extracting electronic health record data or using manual entry and using the UDI-DI to auto-populate structured device data. The usability of the data in AccessGUDID will be evaluated to ensure that this core data is meeting the goal of providing standard device identification data that can be used to inform regulatory decision making. Please see online supplemental file 4 for further clarification on these concepts and a full list of acronyms used in this manuscript.

Request reuse permission