CTSI Annual Pilot Awards to Improve the Conduct of Research

An Open Proposal Opportunity

Overcoming hurdles to patient-centered outcomes research

Proposal Status: 

The Patient Centered Outcomes Research Institute is charged with facilitating informed choice by funding initiatives that create condition-specific registries designed to provide information necessary for patients to understand risks and expected benefits in terms of meaningful outcomes. Traditionally, registries require hiring personnel, not directly involved in the care of patients, to distribute surveys, review charts, fill in forms, and enter data into an offsite registry. We employed a method to generate registry data for joint replacement and spine surgery that uses existing data. All hospitals, clinics and emergency departments use ICD-9 diagnosis, procedure and CPT codes to describe every patient encounter. Unlike the electronic health record, these data are stored in a uniform format. The integrated claims repository employs an ontology, which is a set of terms commonly used by patients and physicians to describe patient demographics, etiology, comorbidities, procedures and outcomes such as complications, readmission and reoperation rates. The ontology script, adaptable to any field of medicine, includes parameter names and codes associate with the parameters. A program reads the script and generates the program that creates the integrated claims repository. The University of California Office of the President recently approved funding create a spine registry by replicating the integrated claims repository at the other UC medical school campuses. The UCSF Committee on Human Research (CHR) approved the project and Notice of Intent to Rely (NOITR) is streamlining approval at the other campuses. The ontology has also been adapted to query national and statewide inpatient and ambulatory datasets.

The first aim of this proposalis to establish a CTSI clinical registry consultancy to help researchers create ontologies to define integrated claims repositories and condition-specific registries. This consultancy will enable collaborators in other clinical fields to create preliminary data for PCORI grants. Steps for generating these data: 1) develop and validate ontology definitions through chart review and queries of the University HealthSystem Consortium database. 2) Define inclusion and exclusion criteria and parameters for the registry. 3) Gain CHR approval with NOITR. 4) Implement the integrated claims repository. 5) Create reports. 6) Replicate this work at other campuses and compile collaborative registry data. 7) Generate state- and nationwide statistics for the condition. Metrics of success are the number of publications and grant submissions arising from claims repositories established in other fields of medicine. This model will become self-sustaining through consultant fees and grant collaborations.

Regulatory issues being addressed include: Determining whether the registry activity is research or a quality improvement activity. What activities require patient consent and HIPAA authorization? What options are available for Business Associate Agreements? What are the concerns of the Privacy Office? What computer security controls are needed? How can data be added or extracted from the electronic health record? What are the restrictions when using cost data? How do you protect sensitive patient information when using iPads or iPhones for data collection?

The second aim of this proposal is to work closely with the CHR and Regulatory and Knowledge Support Director to 1) gain approval to set up integrated claims repositories at the 5 UC campuses, and 2) develop libraries of CHR and consent form templates, business associate agreements, and other regulatory documents. By the end of the pilot, a library of template regulatory documents will be created that will enable student and post-graduate trainees to produce notable results during a well-planned research rotation.

Budget: $70,000. The budget will provide 20% time to refine regulatory documents and establish the registry model beyond spine surgery, and support in the CHR and RKS offices to develop and maintain document template library.

Collaborators: Elizabeth Boyd, John Heldens


I believe this registry could potentially be used as a resource database for my evidence-based sample-size calculation tool. Your excellent list of outcomes relevant to spine research demonstrates the breadth of data needed within a given field and the importance of collaboration with subject-matter experts such as yourself to identify these. My plan it to create access to a non-specialized database and supplement that, as needed.

Thanks Joan, I can show you how to create a flexible script to select cases and define parameters from the National Inpatient Sample. Unfortunately, the only outcomes in this database are complications rather than the longitudinal outcomes available in Medicare, OSHPD or private payer claims. But you can stratify hospitals and physicians by case volume, type of facility and US region.

I like this idea. The Consultation Services (CS) Data Management Unit and Design Unit are already engaged in helping researchers query existing administrative and EMR-based databases to generate patient samples for research purposes; a "clinical registry consultant" could extend this service in a very useful way. Overcoming the regulatory hurdles and facilitating cross-campus registries could really make a difference in helping clinical researchers answer clinically relevant questions. CS would be happy to help develop a consultation unit or set of "niche" consultants to support this.

Thanks Mark. The Stata advice that you provided was instrumental in streamlining the code that we now use to automate the querying of the HCUP and decision support datasets.

I think the idea of establishing a clinical registry consultation service within Consultation Services is a good one that fits well with their overall team. RKS would be interested in partnering on the establishment of a library of consent form templates and other documents. I do think that this is probably more time-consuming than a 20% effort -- I would urge the PI to consider other aspects of completing the project -- for instance, a research assistant, the time of CHR staff to work on these templates, any data resources, etc.

This is a good idea, and the support would help HRPP standardize and streamline practices at UCSF. We should be able to leverage the work performed on this project to improve system-wide efforts. HRPP would be happy to collaborate on this.

Commenting is closed.