Development of an International, Multicenter, Hyperbaric Oxygen Treatment Registry and Research Consortium: Protocol for Outcome Data Collection and Analysis

Background Hyperbaric oxygen (HBO2)—oxygen at pressures higher than atmospheric—is approved for 14 indications by the Undersea and Hyperbaric Medical Society. HBO2’s main effect is to increase oxygen content in plasma and body tissues, which can counteract hypoxia or ischemia. Laboratory studies show that HBO2 has effects beyond relieving hypoxia (eg, promoting angiogenesis in irradiated tissue, anti-inflammatory effects, radiosensitization of tumors, hypoxia preconditioning, and fungal growth inhibition) and has potential to treat conditions such as inflammatory bowel disease and pyoderma gangrenosum. Lack of consistently collected outcome data on a large cohort of individuals receiving HBO2 therapy limits its use for both established and new indications. A course of therapy often involves 30-40 visits to a hyperbaric chamber, so the number of patients seen at any given center is constrained by chamber capacity. As a result, published HBO2 outcome data tend to be from small case series because few patients with a particular condition are treated at a given center. To solve this problem, a registry that collects and pools data systematically from multiple institutions has been established. Objective The aim of this study is to collect consistent outcome data across multiple hyperbaric centers to assess treatment effectiveness and establish a research consortium. Methods A consortium of hyperbaric centers who have agreed to collect consistent outcome data on all patients seen has been assembled. Data are collected at each participating center using Research Electronic Data Capture (REDCap), a web-based, data collection system used frequently for research. Standard outcome measures have been defined for each condition, which are programmed into the REDCap data collection templates. Governance is through a consortium agreement that defines data security, data sharing, publications, liability, and other issues. Centers obtain Institutional Review Board (IRB) and ethics approval to participate, either from their own institutions or by relying on the IRB at the coordinating center at Dartmouth College. Dissemination will occur through a yearly report and by publications based on the data in the registry. Results Early results from some common indications show significant pretreatment to posttreatment changes. Additional indications and outcome measures are being added using the procedures outlined in the consortium agreement. Conclusions The registry collects consistent outcome information for a therapy that needs further study and a stronger evidence base. It also overcomes the challenge of collecting data from an adequate number of patients for both established and emerging indications by combining data collection from multiple centers. The data entry requirements should be within the capabilities of existing staff at any given hyperbaric center. By using REDCap, the registry can be expanded to include detailed information on particular indications and long-term follow-up on selected patients without significantly increasing the basic data entry requirements. Through the registry, a network of enrolled hyperbaric centers has been established that provides the basis for a clinical trial network. International Registered Report Identifier (IRRID) DERR1-10.2196/18857


Introduction
Hyperbaric oxygen (HBO 2 ) treatment, defined as breathing 100% oxygen at pressures greater than 1.4 atmospheres absolute, is used for 14 indications approved by the Undersea and Hyperbaric Medical Society (UHMS), such as soft tissue radiation injury and enhancement of healing in selected problem wounds (see Table 1) [1]. HBO 2 greatly increases the amount of oxygen dissolved in plasma and is effective at relieving hypoxia. This effect is useful in conditions such as ischemic wounds or compromised flaps or grafts, where inadequate oxygenation is a factor. Laboratory studies also suggest HBO 2 has actions beyond the relief of hypoxia (eg, promoting angiogenesis in irradiated tissue, anti-inflammatory effects, radiosensitization of tumors, hypoxia preconditioning, and fungal growth inhibition) and has potential to treat other currently unapproved conditions, such as inflammatory bowel disease and pyoderma gangrenosum [2][3][4]. Approximately 1350 hyperbaric chamber facilities exist in the United States, and outpatient facility claims for hyperbaric services to Medicare alone totaled US $178 million in 2015. Although HBO 2 treatment is used for a variety of indications, much of the evidence to support its use is based on small trials, case series, and retrospective studies (see Table 1) [2,[5][6][7][8][9][10][11].
This inconsistent evidence base has led to a range of opinions about when and how HBO 2 should be used. For example, a common application of HBO 2 is in treating radiation cystitis, which is supported by most insurance policies in the United States [12,13]. The evidence base and practice patterns are strong enough that when a randomized trial of HBO 2 for radiation cystitis was attempted (ie, ClinicalTrials.gov Identifier: NCT00134628), the trial had to be closed due to poor recruitment. It was difficult to find patients who were willing to be randomized to placebo treatment and providers who were willing to refer them. In one review describing the many treatment options for radiation cystitis, the authors concluded "there are currently no adequate treatment options." They cited HBO 2 response rates between 27% and 92%, and recurrence rates after treatment from 8% to 63% [14]. The lower end of this response rate range suggests HBO 2 should only be tried occasionally for selected patients, while the higher end argues that HBO 2 should be the treatment of choice. The largest prospective study of HBO 2 treatment for radiation cystitis included in that review was based on 40 patients [12]. Since that review, five Nordic university hospitals were able to complete a randomized trial for radiation cystitis, although they excluded from enrollment patients with severe ongoing bleeding (ie, the patients where the impact from HBO 2 treatment would be most meaningful). The trial showed benefit from HBO 2 [15].
This diversity of opinion and practice creates a difficult situation where the published evidence base is small, but advancing to large-scale clinical trials has been difficult. Also, hyperbaric centers worldwide are being asked to provide a stronger level of evidence to support the treatments they deliver. For example, the National Health Service (NHS) England will now support the routine use of HBO 2 treatment only for decompression illness and gas embolism [16]. Conditions reviewed and not recommended for routine hyperbaric therapy because of the perceived lack of reliable outcome data include the UHMS-approved indications of carbon monoxide poisoning, soft tissue radiation damage, and necrotizing soft tissue infections. A new approach is needed to collect outcome data for HBO 2 treatments.
The study of HBO 2 treatment presents unique challenges. Because HBO 2 treatment usually requires daily treatments over the course of 1-2 months in a hyperbaric chamber, single hyperbaric centers typically do not treat large numbers of patients and cannot accrue sufficient numbers to conduct credible studies. Furthermore, because patients are referred for treatment from other specialties, follow-up tends to be conducted by the referring specialist, and outcome data are not available to the hyperbaric program beyond the treatment period. Obtaining meaningful data on HBO 2 outcomes requires pooling of data from multiple centers and establishing an infrastructure of centers motivated to conduct research and initiate long-term follow-up. The HBO 2 Registry Consortium described here will provide this framework for urgently needed evaluative observational studies, with the potential to improve the clinical application of HBO 2 dramatically. In addition, the consortium will provide an efficient platform for conducting trials on a wide range of HBO 2 indications, as well as studies of the molecular underpinnings of the treatment itself. This consortium infrastructure could be used to develop the platform from which multiple studies could be conducted at much lower cost. Table 1. Indications for hyperbaric oxygen (HBO 2 ) therapy with Undersea and Hyperbaric Medical Society (UHMS) assessment of the quality and strength of evidence from the UHMS Indications report [1].
Notes relevant to the registry UHMS assessment of evidence:quality b , strength c Indication a

UHMS-approved indications
Used only at certain centers; need to combine data from centers

Non-UHMS-approved indications
No randomized controlled trials for any treatment modality; recommended as second-line therapy [5] Not rated Calciphylaxis Case series show benefit [6,7]  Not rated Ulcerative colitis a Many of these indications are only seen episodically at any given center, so a registry is important for aggregating a sufficient number of cases to draw conclusions. b Quality of evidence has five levels-Level A: highest quality, where evidence comes from more than one randomized controlled trial, a meta-analysis of high-quality randomized controlled trials, or one or more randomized controlled trials corroborated by high-quality registry studies; Level B-R: evidence comes from randomized trials; Level B-NR: evidence comes from nonrandomized trials; Level C-LD: evidence comes from limited data; and Level C-EO: evidence comes from expert opinion. Beyond the fundamental question of whether HBO 2 treatment should be recommended for given indications, the HBO 2 community needs data to answer more detailed questions, such as: Are some forms of radiation injury (eg, brain radionecrosis) more or less responsive than others to HBO 2 treatment? Are some patients more likely than others to benefit from HBO 2 for a given indication, and can those patients be identified? Having identified the patients most likely to benefit from treatment, can trials be designed more effectively to test HBO 2 treatment (eg, crossover trials)? A well-designed registry can provide the data required to answer these questions.

Overview
This paper outlines the development of an international, multicenter, prospective registry consortium. A center joins the consortium by signing the consortium agreement. This agreement covers membership, governance, data sharing requirements, use of member data, publications, intellectual property, liability, confidentiality, and insurance. Data are collected using Research Electronic Data Capture (REDCap), a widely available, easily accessible, Health Insurance Portability and Accountability Act (HIPAA)-compliant, web-based data collection system [17]. Each participating center determines its own start date for data collection (ie, reference date) and records baseline data for every patient referred to the treatment center, whether or not treatment is indicated for, or accepted by, the patient after that start date. Patients included in the registry are those patients who have been evaluated for possible treatment of any UHMS-approved condition, or any non-UHMS-approved condition, including those who are part of research studies or trials. The registry gathers data on whether the evaluation determined that treatment was contraindicated, indicated and scheduled, or indicated but declined by the patient.
Each center's participation is overseen by its own US Institutional Review Board (IRB) or the country-specific equivalent, or a center can opt to rely on the Dartmouth Committee for the Protection of Human Subjects (ie, the Dartmouth IRB). The REDCap-based data collection template is the same at all centers, so all centers collect the same outcome measures. Quarterly, each center performs a deidentified data download from their REDCap database to the coordinating center's REDCap database at Dartmouth College. These deidentified data from each center are combined into a single REDCap database, which is the multicenter data registry. Optional consent for longer-term follow-up is being pilot tested at one center.

Data Collected
The design approach to the registry is to create a system that can be used at any hyperbaric center. This means the system must be low cost and not require excessive staff effort. Although gathering data from electronic medical record systems is desirable to avoid repeated data entry, this is not practical for this project due to the diversity of medical record systems and the level of effort and funding required to standardize and update outcome measures and procedures among them. To minimize staff effort, data entry needs to be minimal, which means extensive data on comorbid conditions, medications, and medical history cannot be collected. Instead, the registry has to focus on a few key outcome measures whose data any center can collect and enter reliably. For studies using registry data where more information on the individual patients is needed, an interested investigator can obtain IRB or ethics approval to work with more detailed data at the individual centers using procedures outlined in the consortium agreement.
The registry database was initially designed by a consensus of founding members of the HBO 2 research consortium-the Geisel School of Medicine at Dartmouth in Hanover, New Hampshire, and the Dartmouth-Hitchcock Medical Center in Lebanon, New Hampshire-and core data are collected for UHMS-approved and some nonapproved conditions. Data are collected using four main data collection instruments (see Table 2). The Demographics instrument collects demographic information as well as information on insurance and distance traveled. The Pre Treatment Information instrument collects information on the condition being treated, treatments prescribed and administered, and subjective and objective measurements of the patient's status at treatment start. A quality-of-life measure-the EuroQol, 5-dimension, 5-level (EQ-5D-5L)-is also administered at baseline on all patients [18,19] as part of the Pre Treatment Information instrument. The Pre Surgical Information instrument collects subjective and objective measures of the patient's status prior to a surgical intervention (eg, tooth extraction) if one is performed. The Treatment and Outcomes instrument repeats information collection of the Pre Treatment Information outcome measures, records the actual treatment given, asks about complications, and repeats the quality-of-life questionnaire.
Both indication-specific outcomes and general outcomes are collected, including HBO 2 treatment complications (eg, changes in refraction, seizures, pneumothorax, confinement anxiety, barotrauma, and placement of pressure-equalization tubes in the ear). Wherever possible, common objective outcome measures are used.  [20,21] as well as some questionnaires that were custom developed for the registry (see Table 2) [22][23][24]. Factors that might affect the effectiveness of the therapy are also collected (eg, diabetes, smoking, and other nicotine use). Future plans, including long-term follow-up and linkage to cancer registries and vital status data, depend on future funding. Table 3 lists the parameters measured [25,26]. Table 2. Data collection instruments and questionnaires used in the registry. Patient-reported outcomes are used in the registry whenever possible.

Data collection instruments
Age, race, ethnicity, biological sex, insurance, and driving distance; personal health information is only kept at individual centers Demographics Referral reason, urgency, diabetes, smoking, nicotine, indication, baseline questionnaires and outcome information, EQ-5D-5L b , and prescribed treatment

Pre Treatment Information
Outcome measures prior to intervention if one is performed (eg, tooth extraction, mandibular reconstruction, etc) Pre Surgical Information Treatment given, complications experienced, and outcome measures Treatment and Outcomes

Questionnaires used in the registry
Seven questions from the Urinary Distress Inventory 6 plus one custom registry question on urinary bleeding [24] Bladder Questionnaire (radiation cystitis) Nine questions custom developed for the registry Bowel Symptoms Questionnaire 37 questions selected from the EORTC c QLQ-H&N35 d [23] and the GRIX e questionnaire [22] Head and Neck Questionnaire Two questions based on the Chandler [20] and RTOG f scales [21] Laryngeal Soft Tissue Radionecrosis Questionnaire 11 questions custom developed for the registry Perianal Crohn's Symptom Index a Details of data collection instruments include the types of data collected; details of questionnaires used in the study; and the sources used to develop the questionnaires. b EQ-5D-5L: EuroQol, 5-dimension, 5-level.  Table 3. Indications for hyperbaric oxygen (HBO 2 ) treatment and outcomes measured for the registry.

Assessment of cyanosis in affected areas pre-and posttreatment and amputations posttreatment
Acute ischemia (not crush injury or compartment syndrome) Number of wounds and wound measurements Acute thermal burn injury Glasgow Coma Scale score for brain events, troponin for cardiac events, and six-level outcome measure for all

Governance
The HBO 2 Registry Consortium includes member institutions fulfilling the requirements as shown in Textbox 1. A steering committee is responsible for the governance of the registry. The Geisel School of Medicine at Dartmouth and the Dartmouth-Hitchcock Medical Center are the founding members of the consortium and each have representatives on the steering committee. The steering committee also includes a representative from each participating center. By a majority vote, with affirmative votes from the founding members, the steering committee has the authority to set the strategic direction for the registry, admit or remove members, set membership fees, establish policies, and relocate the registry. Member-initiated research protocols are approved by a majority vote of the steering committee. The committee also votes on any changes to the REDCap data collection instruments. The steering committee meets annually at the UHMS annual meeting, where changes and modifications to the registry are discussed and voted upon. Template IRB application materials are freely shared between participating centers. Textbox 1. Requirements and activities of participating centers of the HBO2 Registry Consortium.
Center characteristic: • Hyperbaric oxygen (HBO 2 ) treatment program Administrative requirements: • Ethics Committee approval, either by relying on Dartmouth or at own center • Execution of consortium agreement, which includes agreement to share deidentified data with coordinating center and willingness to have data used for registry purposes

Statistical Analysis
Descriptive analyses will be conducted to describe the baseline characteristics and outcomes of patients with each indication, overall and by center, including the numbers, proportions, and confidence intervals for the following: patients with each indication for treatment; those evaluated who were not treated or who were partially treated and the reasons why; baseline characteristics, including referral information and baseline outcome measures; treatments given; occurrence of each type of side effect; and outcome measures at the end of treatment with change scores and relative change scores, as appropriate.
For measures with numerical scores, we will assess absolute and relative changes; for other measures, we will report outcomes in terms of the proportions with symptom resolution, improvement, no change, or deterioration. Factors associated with key outcomes will be explored using multifactorial analyses where sample size is sufficient for models to be stable. All results will be presented and discussed in terms of estimates and 95% confidence intervals to avoid focusing on P values alone.

Results
Data collection started within the registry at the Dartmouth-Hitchcock Medical Center as a pilot project in 2012. During this time, the consortium agreement was developed and processes were established to enroll other centers. In 2019, four additional centers began entering patient data. Currently, the registry has 919 individual patient entries from four centers, with the majority from the Dartmouth-Hitchcock Medical Center (n=621). Over time, the registry is gradually collecting sufficient data to robustly explore changes before and after HBO 2 treatment. Figure 1, for example, shows the results for the Head and Neck Questionnaire before and after HBO 2 treatment. Figure  2 shows changes in patient-reported symptoms of xerostomia (ie, dry mouth) after receiving HBO 2 treatment. These responses are important because whether HBO 2 treatment has an effect on xerostomia has been a longstanding question in the HBO 2 field [27]. Figure 3 shows changes in the hematuria score for those patients undergoing treatment with HBO 2 for radiation cystitis. Figure 4 shows the experience to date for patients treated for idiopathic sudden sensorineural hearing loss. These preliminary results show that the registry is useful for tracking trends in outcomes and patient-reported symptoms after HBO 2 treatment. As further patient experiences are collected, these data will also be used in retrospective analyses to determine the characteristics of those patients who responded well and of those who did not. Figure 1. Delayed radiation injury. Scores on the Head and Neck Questionnaire before and after hyperbaric oxygen (HBO 2 ) are shown: lower scores indicate fewer symptoms. This questionnaire is administered to any patient who had experienced head and neck radiation and is being treated for radiation injury in the head and neck region. Although responses vary between patients, results show lower scores posttreatment (16.9 pretreatment to 14.3 posttreatment, P=.03, Wilcoxon signed-rank test). By identifying patients who did not respond or who worsened, these data can guide further analyses.  . Delayed radiation injury, radiation cystitis. Scores on the hematuria scale before and after hyperbaric oxygen (HBO 2 ) treatment are shown: 0=no hematuria, 1=microscopic hematuria, 2=occasional macroscopic hematuria, 3=frequent macroscopic hematuria, and 4=severe hemorrhagic cystitis.
Most patients see an improvement in hematuria score (2.5 pretreatment to 1.0 posttreatment, P<.001, Wilcoxon signed-rank test). As the number of entries in the registry grows, these data may be useful for assessing the number of treatments needed for successful outcomes. . As the number of cases in the registry grows, these data could be used to assess how long after the hearing loss HBO 2 may be useful.

Principal Findings
An HBO 2 treatment outcomes registry is feasible and can provide consistently recorded outcomes from HBO 2 treatments at multiple centers. Initial analyses of some of the outcome data are already showing significant changes after HBO 2 treatment.

Strengths of the Registry
A key strength of the registry is its use of REDCap [17], a free, secure, web-based, data collection system used to build and manage online surveys and databases in more than 3964 centers worldwide. The HBO 2 REDCap database has been pilot tested and revised at the coordinating center for 6 years, and core variables have been collected on UHMS-approved indications for HBO 2 treatment, as well as some non-UHMS-approved indications. The database includes 1996 variables, many of which are disease specific and are programmed to appear only for particular indications or situations (ie, during data entry, a staff member will only be entering data on a small subset of the variables available). Data entry takes an average of 15 minutes per case, distributed over several clinical visits. The template for the database can be exported to Excel, emailed to another center, and easily uploaded to create an identical registry that is ready for data entry. REDCap provides the ability to perform a deidentified data export from each participating center. This is sent to a central, pooled REDCap database at the Geisel School of Medicine at Dartmouth. Because of the simplicity of the software and the relatively low burden of data entry, the registry is relatively inexpensive for new centers to install and maintain. Because only deidentified data are pooled, the risk to participants is minimal.

Limitations of the Registry
At present, the registry is following individuals who receive HBO 2 treatment, and a registry of similar patients with similar conditions who are not receiving HBO 2 treatment does not exist.
Therefore, there is no way to compare outcomes from HBO 2 treatment directly with outcomes from similar patients who did not receive the treatment. As the registry grows, an additional limitation will be funding, which will be needed to maintain a system of data-quality oversight, analysis, longer-term follow-up, and other registry-related activities.

Conclusions
An outcome-focused registry for HBO 2 treatment is needed urgently to provide both patients and providers with the information they need to decide whether and how to use HBO 2 treatment. To be successful, this registry must be practical, easy to use, and easily expandable. The registry described here meets these requirements. The data entry requirements are not excessive and should be within the capabilities of existing staff at any given hyperbaric center. Also, by using REDCap, the registry can be expanded to include detailed information on particular indications and long-term follow-up on selected patients without significantly increasing the basic data entry requirements.