| subtitle | title |
|---|---|
Author 1, Author 2, … , Corresponding Author |
Patient Engagement in Clinical Data Research Networks: The 2015 Greater
Plains Collaborative Health and Medical Research Family Survey
|
JAMIA (pretty high bar & impact, but it has already published a lot on PCORnet & CDRNs (see bibliography). I believe they would be positive about describing governance, multi-center data sharing, redcap, etc. I believe they would also be much less critical about the extremely low response rate. & so we therefore can do less with the data/results itself.)
Other Journals???
Describe patient engagement findings and experiences of Healthy Weight Cohort GPC member sites
Placeholder for abstract text. [Larry]
Background: In 2013, the Patient Centered Outcomes Research Institute (PCORI) funded 11 Clinical Data Research Networks (CDRNs) to form the backbone of PCORnet, the National Patient Centered Clinical Research Network. Based on electronic health records (EHRs), each CDRN committed to building large patient cohorts with electronic clinical data governed by policies and procedures to conduct multi-site clinical effectiveness research. The Greater Plains Collaborative (GPC) CDRN consists of 11 academic medical centers in 8 Midwestern states.
Objective: PCORI Phase 1 funding directed CDRNs to demonstrate the feasibility of patient contact and engagement using the EHRs data infrastructure. We describe the GPC feasibility study to engage patients on health and medical research.
Methods: governance, extracts, survey, analysis methods,
Results: Governance & process findings? Summarize tabular results.
Conclusions: Comments on results. CDRNs are a rapid and effective method to assemble large patient cohorts and conduct multisite, patient centered studies of clinical comparative effectiveness research.
In 2015, The Greater Plains Collaborative was established as Clinical Data Research Network (CDRN), funded by the Patient-Centered Outcomes Research Institute (PCORI) to securely collect and organize patient health information obtained during routine care in its member health systems {about PCORnet}. To date, 13 such CDRNS have been funded, creating a national “network of networks”. These networks are organized by a coordinating center and overseen by the National Patient-Centered Clinical Research Network (PCORnet).
The purpose of CDRNs and PCORnet is to support efficient clinical research by creating centralized access to the de-identified data of millions of patients across the country. Each CDRN is responsible for harmonizing patient data across its member systems, and for creating streamlined governance and procedures to facilitate researcher access. Importantly, CDRNs actively involve a variety of stakeholders, including patients, clinicians, healthcare system leaders, and others to build and oversee CDRN activities
To test each CDRN’s ability to identify and recruit patients with a particular condition, and to test the ability to harmonize data elements within a network, each CDRN was required to create three cohorts: one of a common disease, one of a rare disease and one concerning height and weight.
Placeholder text, map graphic?
In Phase I of funding, the Greater Plains Collaborative (GPC) was comprised of ten member health systems, with the data of roughly 6 million people across 7 states, north and south across the Great Plains region. Member institutions included University of Kansas Medical Center, Children’s Mercy Hospital (Kansas City, MO), University of Iowa Healthcare, University of Wisconsin - Madison, Medical College of Wisconsin, Marshfield Clinic, University of Minnesota, University of Nebraska Medical Center, University of Texas Health Science Center at San Antonio, University of Texas Southwestern Medical Center.
Covering more than 1300 miles, the broad reach of the GPC network encompasses large swaths of rural populations as well as multiple urban centers. Four systems in the GPC have established significant relationships with Native American populations. Two health systems located in Texas serve heavily Spanish-speaking populations, which had a significant effect on the healthy weight study approach and methods. Of the ten member health systems, all provide comprehensive adult and pediatric care, with the exception of Children’s Mercy Kansas City, which exclusively serves children.
In Phase II, largely following the work described in this paper, two additional members were added – University of Missouri and Indiana University.
The healthy weight cohort team began regular meetings in January of 2014. Based on discussion and collective interest, the group quickly decided to develop its cohort and survey around a pediatric population. Weekly working group calls established an interest in characterizing the cohort around data elements that would be attainable for the nascent GPC network, and creating a survey that could be used as a building block for future GPC and healthy weight cohort work. Thus, the Health and Medical Research Family Survey focused on respondents’ willingness to take part in research.
The overall purpose of the Greater Plains Collaborative (GPC) Healthy Weight Study (HWS) was to conduct a demonstration survey across the 12 participating GPC sites focused on the topic of pediatric obesity. More specifically, the aims were to: 1) estimate the willingness of individuals to be contacted about research activities; 2) obtain information on the attitudes of parents and adults of child bearing age about research, including participation of their child/ren; 3) gain insight into participant attitudes about the use of gathered data for both local and national research; 4) explore the impact of various demographic factors, survey methods, regional variation, and weight status on the above questions. An additional expected outcome of the project was to establish and explore the various practicalities and functionalities of a large semi-interconnected system such as the GPC for conducting collaborative research focused on pediatric obesity.
Although previous studies have been published on adult obesity using a PCORI funded CDRN (Young et al, 2016), these studies were retrospective in nature reporting on the number of patients in the network who met certain criteria. In contrast, the GPC HWS not only gathered retrospective data on individuals who met certain specific inclusion criteria but also contacted a random group of individuals from this sub-sample at each site with a survey invitation. The goal for of the 12 participating sites was to collect 1000 responses to “Survey 1” and 100 responses to “Survey 2.” Survey 1 was a single question asking whether or not the individual contact was willing to be contacted for research, and survey 2 was a more detailed assessment (see Table 1).
Across the GPC a central IRB process was developed (asked Russ for reference on this – not in his paper). This involved a legal agreement between all sites designating the University of Kansas Medical Center as the lead IRB for the GPC, with all other sites agreeing to “rely” on this central IRB. For the healthy weight study this central IRB process was utilized. However, the University of Texas Health Science Center at San Antonio (UTHSCSA) served as the central IRB site for the healthy weight study across the GPC network. A team of investigators and staff developed the necessary IRB documents which were submitted to the IRB at UTHSCSA. Once the documents were reviewed and approved by this single IRB, the documents were shared with all other participating site’s IRBs, and these documents were approved without additional edits under the existing overall IRB reliance agreement.
Individuals who were sent the survey invitation were contacted via a variety of means, including snail mail, email, and through the patient portal systems in the site’s electronic medical record (EMR). The method of contact was selected by the site as that which would encounter the fewest barriers and result in the greatest sample. Allowing each site to select their own contact method allowed for a naturalistic experiment of response rates and other factors by method of contact. For a detailed list of site and contact method see Table 2.
This study combined electronic chart review with the results of a survey administered to patients, or in the case of minors, to the patients’ guardians. For the electronic chart review, data was extracted from an open source data warehouse platform called Integrating Informatics from Bench to Bedside (i2b2) [@murphy_instrumenting_2009]. Each of the participating sites had an i2b2 instance deployed, where they stored a de-identified version of all structured data from their respective EMR systems. As a result, no site had to transmit any identifying information to any other site. Though not all the sites used the same type of EMR system, because of i2b2’s very flexible star-schema design the task of combining data from disparate sources was made more tractable. The patient surveys were administered via REDCap [@reference_forthcoming]. The survey data and the EMR data extracted from i2b2 could be merged together because each record is keyed to the same non-informative index accross the two data sources.
Cohort selection was also done using i2b2, using the selection criteria shown in Table 3.
There was one supplementary request sent out for the study sites to acquire from their EMRs information about race, ethnicity, and class of insurance provider (private, Medicaid, employer, government, self-pay, etc.). Together with these data elements, information about income was requested. Since not all health systems record patient income, we had to rely on median household income for the census block group in which each patient’s address was located, as obtained from the 2013 American Community Survey [@reference_forthcoming].
Data was extracted from the EMR using DatBuilder [@reference_forthcoming] collated into an analyzable tabular form using DataFinisher [@bokov_denormalize_2016].
Placeholder text. [Larry with assistance from Alex]
Tables 1…N, Figures 1…N – describe findings [Larry-to write describe, Alex to produce table outputs]
conclusions, strengths, limitations, future studies [Larry – would like assistance from everyone….]
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum
Several potential additional references are here:
https://academic.oup.com/jamia/search-results?page=1&q=PCORI&SearchSourceType=1
[1]
R. L. Fleurence, L. H. Curtis, R. M. Califf, R. Platt, J. V. Selby, and J. S. Brown, “Launching PCORnet, a national patient-centered clinical research network,” J Am Med Inform Assoc, vol. 21, no. 4, pp. 578–582, Jul. 2014 [Online]. Available: https://academic.oup.com/jamia/article/21/4/578/2909226/Launching-PCORnet-a-national-patient-centered. [Accessed: 12-Apr-2017]
L. R. Waitman, L. S. Aaronson, P. M. Nadkarni, D. W. Connolly, and J. R. Campbell, “The Greater Plains Collaborative: a PCORnet Clinical Research Data Network,” J Am Med Inform Assoc, vol. 21, no. 4, pp. 637–641, Jul. 2014 [Online]. Available: https://academic.oup.com/jamia/article/21/4/637/2909307/The-Greater-Plains-Collaborative-a-PCORnet. [Accessed: 12-Apr-2017]
Table or figure caption.
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.
Table 1. Survey 1 and survey 2
Table 2. List of each site with their method of contacting survey participants.
Table X. format
Tables 2,3,4 – formats - weighted and unweighted modeling