Comprehensive Landscape Analysis for Usable Real-World Wound Care Data

© 2024 HMP Global. All Rights Reserved.
Any views and opinions expressed are those of the author(s) and/or participants and do not necessarily reflect the views, policy, or position of Wounds or HMP Global, their employees, and affiliates.

According to a 2023 published analysis of 2014-2019 Medicare claims data, 16.4% of Medicare beneficiaries (10.5 million people) were affected by chronic wounds in that period, the annual cost of which was conservatively estimated at $22.5 billion but may have been in excess of $67 billion¹; this is a significant economic impact.² RCTs, which occupy the apex of the evidence pyramid, are designed to demonstrate the efficacy of a treatment under ideal conditions. These studies look at a particular wound type for a single treatment period, usually 12 to 16 weeks, and a narrowly defined population with many comorbid conditions excluded. RCTs provide limited insights into the natural history of specific wound types, their incidence in the general population, the number of concomitant wounds per patient, locations of care, wound outcome, frequency or number of recurrences, and the association of comorbid conditions with failure to heal.^3-5 This information can be obtained from RWD.^5-7 RWD are vital for identifying gaps in evidence-based clinical practice, disparities in care, unmet therapeutic needs, and promising signals that a medical treatment may be of benefit to a broader population than was explored in RCTs. 

The FDA is placing increasing value on RWE in regulatory decision-making, and this has increased the demand for RWD in the field of wound care.⁸ To date, analysis of large, structured data sets in this field have been limited to administrative data. Although interoperability standards facilitate the exchange of already structured health care information (eg, diagnosis codes, procedure codes, claims), vital clinical observations such as wound descriptive characteristics (eg, drainage character, tissue type exposed, amount of necrotic material, wound outcome) are typically recorded as unstructured data, if they are captured at all.⁷ Even when the challenges of data reliability are overcome, organizations that hold large repositories of electronic patient information are often reluctant to allow access to it due to well-founded concerns around privacy and security.^9,10 Legal agreements must be implemented between the custodians of RWD and potential researchers, and the significant technical resources required to implement rigorous de-identification protocols prior to data sharing are beyond the capability of most health care institutions. Ambitious projects to create so-called data maps of clinical observations, thus linking diverse sources of data such as the Observational Medical Outcomes Partnership (OMOP),¹¹ are currently beyond the reach of the field of wound care. 

The WCCC is a 501(c)3 nonprofit, FDA-recognized collaborative community focused on stimulating wound care innovation and improving patient access to cutting-edge technology.¹² The WCCC, through its Real World Evidence Work Group, recognized that its organizational goals could not be achieved without access to high-quality RWD. In May 2023, the WCCC undertook a project to identify US sources of high-quality, research-ready RWD on chronic wounds that were available for use. This effort is the foundation for a larger project—the Natural History Project—that is focused on describing the natural history of DFUs and VLUs, and the patients who experience them.

Initial efforts to identify suitable RWD began with creating a set of evaluation criteria pertinent to the Natural History Project and an initial list of over 30 potential wound DBs to profile. The profiling effort was too time and labor intensive for the WCCC, as a volunteer organization. Thus, it was decided to contract with DGI, LLC (owners of B.R.I.D.G.E. TO DATA; hereafter “search engine for health care DBs”), an organization that specializes in these types of projects.¹³

Criteria for DB selection included the availability of longitudinal, already structured, analyzable data on wound characteristics, treatment, and outcomes in chronic wounds from any site of care and across a broad geographic area. Registries were the primary focus, but other DB types were also considered. Wounds of interest included DFUs, VLUs, pressure ulcers, chronic non-pressure ulcers, traumatic wounds, surgical dehiscence, and arterial ulcers. After the contracted company identified potential DBs, WCCC volunteers narrowed them down to those of interest and queries were submitted to the DB custodians regarding available data on the wounds of interest. Each identified DB was evaluated if it was found to be both suitable and usable.^10,14

The authors undertook the following steps to evaluate the possible sources of information (landscape) of chronic wound RWD: (1) searching for potential DBs, (2) profiling DBs for their data attributes, (3) focusing attention on the most suitable candidates, and (4) excluding DBs that are not usable. 

Search strategy

This systematic DB landscape analysis involved multiple data sources and search strategies. Initially, structured searches were conducted by the contracted company in the search engine for health care DBs and PubMed (top 50 results) to assess the feasibility of this project (Supplement Table 1). Experts in wound care (L.G.V., L.G., C.E.F., V.R.D.) reviewed the following 6 factors—DB name, US region, DB type, data source, population type, and wound types—and provided feedback on whether a DB should be included, indicated DB limitations, and suggested priority of interest (low, medium, or high). Next, screening in PubMed was extended, and unstructured, free-text web searches were performed by the contracted company to identify additional DBs, key stakeholders in the field of wound care, or events in the chronic wounds field (eg, high-visibility researchers, conferences, medical centers). Key stakeholders and conference organizers were contacted for more information about potential DBs. Field experts and professional contacts were also consulted for knowledge of additional DBs.

Data compilation

Supplement Table 2 lists the chronic wound care data elements of interest in the RWD search, based on the authors’ experience and previously published work related to applicability of RCTs to general wound care populations.^4,5 A comprehensive table of all DB search results was generated, and as many data fields as possible were verified by contacting DB managers. Data dictionaries, published literature, and other online information was also reviewed. The table was based on the methodology developed by the company that owns the search engine for health care DBs, allowing for side-by-side comparisons of standardized data fields. If a data field was not routinely captured in a specific DB, it was marked as “not available.” Any data field that could not be verified was marked as “not verified yet.” At least 3 attempts were made to have a DB manager verify the data fields.

Screening and selection

DBs were excluded if the data were not available for research, were no longer updated, or if DB manager verification could not be obtained. DBs were labeled as “suboptimal” if they captured some, but not all of the critical data fields. DBs were marked as “not feasible” if they captured critical data but were logistically impractical for use. The remaining DBs were flagged high potential. All identified DBs were initially manually screened by 1 author (A.J.K.) for their relevance and marked for inclusion/exclusion, followed by additional screening by other authors (J.R., L.G., and S.A.K.) and field experts. The reviewers reached a consensus on the final categorization and DB selection. 

The initial feasibility search revealed 29 DBs that met the minimum criteria of systematically capturing data on 1 or more chronic wounds in the US population and being a registry and/or an EHR system (Figure). Three DBs were found not to provide data for research use and were excluded. Combining the initial search results with a second comprehensive multifaceted search yielded a total of 34 DBs. After several communications with respective DB managers, the review team deemed 5 candidate DBs to be of highest potential for the Natural History Project because they captured detailed real-world clinical data and wound care characteristics in US populations; however, 1 of these 5 DBs was currently authorized only for use toward the WCCC’s nonprofit mission but may not be available to the larger research community. Reasons for excluding DBs included but were not limited to the following: DB not capturing critical wound characteristics or clinical data (none [n = 7], suboptimal [n = 5]), DB managers not being responsive (n = 6), nonfeasibility of DB access by the research team (eg, cost, laborious data licensing procedure, data via chart review only) (n = 4), and no established process for data licensing by DB owner (n = 1).

Among the 5 high-potential candidate DBs for the WCCC’s Natural History Project, 4 (80%) had coverage in multiple US states, while all 5 (100%) included data extracted from an EHR system (Table 1). Out of 4 DBs with DB sizes provided, 3 (75%) covered more than 200,000 unique patients with more than 500,000 discrete wounds. All 5 DBs covered data within the past 15 years, while 1 DB started capturing data as far back as 2001. Most EHRs started recording data in 2007, but some data was captured as far back as 2001 although it is unlikely it may be usable. All 5 DBs included outpatient data, while only 2 (40%) included inpatient data. All 5 DBs also captured data on all chronic wound types of interest, general medical and patient data, and prevalence rates of comorbidities; however, 4 DBs (80%) did not record patient data from clinical trials, 1 (20%) indicated that data on patient procedures were limited, and 1 (20%) did not collect payer information. All DBs with high potential (5 of 5 [100%]) were used in published studies on chronic wounds. 

While there was variability in how much detailed wound care data each candidate DB collected, all DBs captured data by wound type (for all wounds at each visit) as well as some or all of the outcomes of interest (Table 1). Three (60%) of the 5 DBs also systematically recorded important data on wound characteristics and tissue type exposed; other interesting methods for acquiring those data included a unique process of linking to any registry or searching physician notes for descriptive wound data.

One of the largest data sets that includes the most vulnerable patients (nursing home residents) did not have a data privacy and data licensing structure in place; therefore, it could not be used by third-party researchers. Another large DB that met the criteria for the Natural History Project and is currently available only to the WCCC may not be suitable or usable for the broader research community.

This project identified 5 DBs with high potential to answer key questions about patients living with wounds, although 1 DB has data access restrictions, but only 1 DB (the US Wound Registry)¹⁵ that is fit for purpose of the Natural History Project. 

The purpose of this effort was to identify usable sources of data that can enable analyses of the real-world journey of chronic wound care patients. 

Engaging in these chronic wound care RWD landscaping steps resulted in the development of a DB Decision Checklist for assessing suitability and usability (Table 2). “Suitability” refers to whether the DB contains data that are fit for purpose and, importantly, ensures the researcher’s understanding of data provenance.^14,16,17 For example, in this process some DBs were excluded for not containing a full clinical picture of the chronic wound care patient (eg, administrative claims DBs), or for not capturing qualitative and quantitative data on wound characteristics (Supplement Table 1). Even after overcoming the hurdle of identifying a DB that captures granular clinical and wound characteristics data, it is not guaranteed that the data are “usable,” that is, accessible, available, and feasible for research use. In the DB landscape project, some DBs were considered impractical due to costly data licensing fees for a nonprofit research group, while others were not accessible because they did not offer their data to external researchers. The WCCC will utilize this information for its Natural History Project; however, the authors of the current report also anticipate that other wound researchers needing RWD will benefit from the DB Decision Checklist and the identification of DBs that are both usable and fit for purpose.

In the United States, it is now mandatory for health care systems to utilize a certified EHR system for the collection of clinical data, yet there are few existing sources of high-quality RWD on chronic wounds. Despite 2 decades of focus on the adoption and use of EHRs, the exchange of structured health care data is limited due to regulatory, legal, technological, and proprietary barriers. Clinical observations are still largely unstructured, unless the EHR system has been purpose built for research. Although the prevalence rate of chronic wounds among Medicare beneficiaries is higher than that of heart failure and treatment of chronic wounds is nearly as expensive as treatment of heart failure in this population,¹ the field of wound care has no comparable investment in research. Because there are almost no pharmacological treatments for chronic wounds, the field has not benefited from investments made by drug manufacturers in the development of RWD. Because few adverse events are associated with wound care treatments, wound care has not benefited from the investments in device safety monitoring in which programs such as the National Evaluation System for health Technology Coordinating Center (NESTcc)¹⁸ receive funding from the FDA’s Medical Device User Fee Amendments (MDUFA)¹⁹ fees. In the United States, unlike in certain other countries (eg, Finland, Sweden), the government has not subsidized registries for research, nor have large wound care research alliances evolved, such as The Australian Health Research Alliance in that country.²⁰ Medicare does not mandate registry data submission for any wound care devices as it has done in other fields, for example, cardiology, where cardiologists who provide implanted defibrillators must submit data to a registry. For most surgeons, participation in a quality and safety registry is mandatory to maintain surgical privileges, but wound care practitioners who perform low-risk, minor procedures have no such requirements. Unlike high-visibility acute care programs (eg, trauma centers), hospital-based wound centers provide primarily outpatient care, and there are no incentives or requirements for these departments to participate in registries, nor are there resources to support efforts that involve secondary data entry. Due to lack of any regulatory, payment, quality, or safety imperatives, the only mechanism currently available for the creation of a large wound care registry is the direct transmission of structured EHR data from a purpose-built EHR system. Some proprietary entities do have their own wound-specific DBs, facilitated by customized EHR systems. 

The lack of a Medicare-mandated quality assurance program that is specific to the field of wound medicine and surgery precludes support for a dedicated wound DB that could be used to promote quality and safety. Using RWD to generate RWE, as compared with clinical trial outcomes in wound care, can aid in (1) understanding the healing trajectory of outliers, that is, those patients who would be excluded from clinical trials due to disease complexity or wounds that do not fit the tight standards for clinical trials; (2) identifying wound characteristics that do not respond to topical treatments despite use according to package labeling; and (3) identifying patients who heal better than expected (this can lead to new studies about what patient characteristics promote healing as opposed to the usual approach of what impedes healing); and it can (4) provide valuable data to the FDA, payers, and legislators that illustrate the complexity of patients with chronic wounds and the true performance of new interventions in real-world patients. 

The WCCC anticipates that data accessed this way will provide valuable insights for the FDA, payers, and legislators, thus illustrating the complexity of patients with chronic wounds and the effectiveness of current and future treatments. 

Despite efforts to identify suitable DBs for RWD analysis, limitations exist in terms of access to certain DBs due to high costs, complex licensing procedures, and lack of data-sharing agreements. Moreover, the reliability of the collected clinical data may vary depending on factors such as data collection methods, potential biases, and adherence to standardized protocols. This study focused primarily on specific types of DBs, potentially excluding valuable data from other sources, such as administrative DBs or clinical trials. Findings may be limited to the DBs analyzed within the study’s scope, limiting their applicability to broader settings or populations. Despite efforts to validate data fields, challenges related to data quality, completeness, and variability in collection methods may persist.

The importance of RWD in health care research is well recognized, but significant challenges persist in accessing comprehensive, high-quality DBs for analyzing chronic wound care. The current study identified a limited number of DBs suitable for the Natural History Project, highlighting the need for health care systems to develop more accessible and usable DBs that capture the full patient journey and wound care outcomes. Moving forward, efforts to address data accessibility, reliability, and comprehensiveness, coupled with advancements in artificial intelligence and regulatory frameworks, are essential to harnessing the full potential of RWD while ensuring patient privacy and safety, and improving outcomes for patients with chronic wounds.