Gang Fang, ETrUOP Research Program
The U.S. is facing great challenges in sustaining its trillion-dollar-cost health-care system, which ranked the highest in spending among all the developed countries, and with its continued annual increases in costs yet unclear improvements in care outcomes. The struggle between constraining surging health-care cost and providing high quality of care has also become a vexing international problem for both developed and developing counties. However, as the Chinese word for “challenge” also means opportunity, this challenge ushers in new avenues for research.
A suggested solution for the challenge is to build an efficient health care with improved quality of care and reduced waste. To improve the efficiency of health care, research is needed to investigate whether health-care resources has been overused or underused as both cause decreased quality of care and increased health-care cost. More importantly, translation of the research into clinical practice imperatively requires the understanding of overuse or underuse or of appropriate use of particular treatments for specific medical conditions in specific populations in terms of patient-centered outcomes (e.g. heart attack, stroke, heart failure, mortality, length of survival, acute exacerbation etc.). We need to answer questions such as these:
- What’s the utilization rate of particular treatments in patient populations?
- How many patients appropriately use a particular treatment?
- Do we have evidence to determine appropriate use or use rate of a particular treatment?
- Do higher treatment utilization rates of particular treatments for a specific condition always lead to higher beneficial outcomes in the population or specific populations?
- Are higher costs associated with the increased treatment utilization rates justified by the higher benefit?
When multiple alternative treatments are available, the question becomes whether a particular treatment is overused or underused versus an alternative treatment, and this circumstance is the comparative effectiveness research (CER) of treatment effects in patient-centered outcomes.
Investigating the real-world treatment effects for CER is critical to address the above questions, which requires the use of observational data such as large health-care utilization databases at the population level in real-world settings. Research using real-world population-level data is critical for several reasons. We will be able to assess the extent and patterns of utilization of a particular treatment in population, which may reveal the magnitude of potential issues or problems. The volume of CER questions means that not all can be feasibly addressed through randomized trials. Observational studies can examine larger populations including subgroups often not adequately represented in trials (e.g. minorities, the elderly, rural residents); and treatment harms may be relatively uncommon and take years to appear, requiring large sample sizes.
Our research program is anchored on
Focused clinical conditions: acute myocardial infarction, diabetes and hypertension
Objectives: 1) To measure and quantify treatment utilization including patterns and variations in populations and by subgroups such as race/ethnicity, age groups, gender, clinical distinct subgroups, and across small geographic areas; 2) To understand determinants and modifiers of treatment use variations.
Current research topics:
- Racial/ethnic, age, gender and rural/urban disparities
- Small-area variation in treatment use
- Medication Adherence
- Population Derived Health Literacy
Objectives: To identify optimal treatment regimen for patient population and subgroups, which will balance the benefit and risk. To provide evidence to support clinical treatment decision and health-care-policy decision to improve the quality of care.
R01 ( 1R01AG046267-01A1, Co-PIs: Bailey & Fang) 09/2014-08/2018
National Institute of Aging $1,653,067 (Total)
National Health Literacy Mapping to Inform Healthcare Policy
Role: Primary PI 20% Effort
Clinical and Translation Science Awards (CTSA) (Lead PI: Runge) 09/2013-04/2018
National Institute of Health
North Carolina Translational & Clinical Sciences Institute (NC TraCS) – CER Strategic Initiative
Role: Co-Investigator 10% Effort
R21 (1R21AG043668-01A1, PI: Fang) 07/2013-06/2015
National Institute on Aging $417,234 (Total)
Should the elderly have lower dose of ACE inhibitors for prevention after AMI?
Role: Principal Investigator 20% Effort
PCORI Pilot Projects (1 IP2 PI000075-01, PI: Stürmer) 9/1/12-12/31/14
Patient-Centered Outcomes Research Institute $673,395 (Total)
Methods to Increase Validity of Comparative Effectiveness Research in the Elderly
Role: Co-Investigator (Lead investigator for Project 1) 20% Effort
DEcIDE-2 RTFO Task Order 3 (HHSA290201000014I, PD: Brookhart) 9/12-7/14 AHRQ $1,112,042 (Total) Developing and Evaluating Methods for Record Linkage and Reducing Bias in Patient Registries Role: Project 1&2 Co-I 20% Effort, and Project 3 PI 20% Effort
National Clinical Research Program (10CRP2610053, PI: Fang) 1/1/10-12/31/11
American Heart Association $110,000 (Total)
Estimate Comparative Treatment Effectiveness between Tissue and Non-Tissue ACE Inhibitors for Myocardial Infarction in the Elderly
Role: Principal Investigator 30% Effort
R01 (HS018381-01A1, PI: Brooks)
AHRQ $1,598,204 (Total)
Comparative Effectiveness of Treatment Combination Post-MI in the Elderly
Role: Co-investigator 20% Effort from 07/01/2010 to 06/30/2011
RC4 (AG038635-01, PI: Brooks)
NIH $1,458,033 (Total)
Interpreting IV Estimates with Treatment Effect Heterogeneity: ACE/ARBs & Race
Role: Co-investigator 20% Effort from 9/01/2010 to 6/30/2011
R21 (HS019574-01, PI: Brooks)
AHRQ $492,261 (Total)
Should High-dose Statin Utilization Rates be Increased for Complex AMI Patients?
Role: Co-investigator 10% Effort from 09/01/2010 to 06/30/2011
Should the elderly have lower dose of ACE inhibitors for prevention after AMI?
Funding: NIA R21 (1R21AG043668-01A1, PI: Fang), Role, PI.
Developing and Evaluating Methods for Record Linkage and Reducing Bias in Patient Registries
Funding:AHRQ RTFO (HHSA290201000014I, PD: Brookhart and Stürmer) Role, Project 3 Lead Investigator.
Methods to Improve Comparative Effectiveness Research in the Elderly
Funding:PCORI Pilot Projects (1 IP2 PI000075-01, PI: Stürmer) Role, Project 1 Lead Investigator
Improving Preference-based Instrumental Variable with Shrinkage Estimation
Racial/ethnic and gender gaps in the use and adherence of evidence-based preventive therapies among elderly Medicare Part D beneficiaries after acute myocardial infarction
Funding:AHA National Clinical Research Program (10CRP2610053, PI: Fang)
Publication: Racial/Ethnic and Gender Gaps in the Use and Adherence of Evidence-Based Preventive Therapies among Elderly Medicare Part D Beneficiaries after Acute Myocardial Infarction. Julie C. Lauffenburger, Jennifer G. Robinson, Christine Oramasionwu, and Gang Fang. Circulation. 2014; 129: 754-763. doi: 10.1161/CIRCULATIONAHA.113.002658
Accompanying Editorial: Not There Yet: Medicare Part D and Elimination of Cardiovascular Medication Usage Socio-Demographic Disparities after Myocardial Infarction. Michelle A. Albert. Circulation. 2014;129:723-724. doi: 10.1161/CIRCULATIONAHA.113.007322
Small-area variation in patient adherence to evidence-based preventive therapies in older adults post acute myocardial infarction
Funding:AHA National Clinical Research Program (10CRP2610053, PI: Fang)
Publication: Prevalent But Moderate Variation Across Small Geographic Regions in Patient Nonadherence to Evidence-based Preventive Therapies in Older Adults After Acute Myocardial Infarction. Fang, Gang PharmD, MS, PhD; Robinson, Jennifer G. MD, MPH; Lauffenburger, Julie PharmD; Roth, Mary T. PharmD, MHS; Brookhart, Maurice Alan PhD. Medical Care. 2014; 52 (3): 185-193. doi: 10.1097/MLR.0000000000000050
Rishi J Desai, Jaya K Rao, Richard A Hansen, Gang Fang, Matthew Maciejewski, Joel Farley. Predictors of Treatment Initiation with Tumor Necrosis Factor-? Inhibitors in Patients with Rheumatoid Arthritis. J Manag Care Pharm. 2014;20(11):1110-20
Rishi J Desai, Jaya K Rao, Richard A Hansen, Gang Fang, Matthew Maciejewski, Joel Farley, Tumor Necrosis Factor-α Inhibitor Treatment and the Risk of Incident Cardiovascular Events in Patients with Early Rheumatoid Arthritis: A Nested Case-control Study. Journal of Rheumatology. 2014 Aug 1. DOI: 10.3899/jrheum.131464
Lauffenburger J, Robinson J, Oramasionwu C, Fang G*, Racial/ethnic disparities in the use and patient adherence of evidence-based therapies post AMI. Circulation. 2014; 129: 754-763
Fang G*, Robinson J, Lauffenburger J, McClurg M, Brookhart M, Regional variation in patient adherence to evidence-based therapies in older adults post acute myocardial infarction, Medical Care. 2014; 52 (3): 185-193
JM Brooks, EA Chrischilles, MB Landrum, KB Wright, G Fang, EP Winer, NL Keating. Survival Implications Associated with Variation in Mastectomy Rates for Early-Staged Breast Cancer.
International Journal of Surgical Oncology 2012; Volume 2012, Article ID 127854, 9 pages: doi:10.1155/2012/127854
Fang G*, Brooks J, and Chrischilles E. Comparison of instrumental variable analysis using a new instrument with risk adjustment methods to reduce confounding by indication. American Journal of Epidemiology. 2012; 38 (11), 1823-1829
Fang G*, Brooks J, and Chrischilles E. Apples and oranges? Interpretations of risk adjustment and instrumental variable estimates of treatment effects using observational data. American Journal of Epidemiology. 2012; 175 (1); 60-65
Fang G*, Brooks J, and Chrischilles E. A New Method to Isolate Local-area Practice Styles in Prescription Use as the Basis for Instrumental Variables in Comparative Effectiveness Research. Medical Care. 2010;48(8):710-717
Brooks J and Fang G. Interpreting treatment-effect estimates with heterogeneity and choice: Simulation model results. Clinical Therapeutics. Apr 2009;31(4):902-919
Fang G*, Zuckerman I, Stuart B, Brooks J. Outcomes associated with the use of thiazolidinediones among medicare beneficiaries with type 2 diabetes – an instrumental variable approach. Value in Health 2007;10(3):A13
Jared M. Freml, Karen B. Farris, Gang Fang, Jay Currie. Recommendations from Iowa Priority’s Brown Bag Medication Reviews: A comparison of student pharmacists and pharmacists. American Journal of Pharmaceutical Education 2004; 68 (2).
Farris KB, Urmie JG, Fang G, Doucette WR, Brooks JM, Klepser D, Fries D and Kuhle C. Population-Based Medication Reviews: A Descriptive Analysis of the Medication Issues Identified in a Medicare Not-For-Profit Prescription Discount Program. Ann Pharmacother. 2004; 38 (11); 1823-1829.
Farris, KB. Kumbera, P. Halterman, T. Fang, G. Outcomes-based pharmacist reimbursement: Reimbursing pharmacists for cognitive services (Part 1 of a 2-part series). Journal of Managed Care Pharmacy. 8(5): p 383-393. 2002.
The research program is equipped with an advanced, dedicated, secured research server with high capacity to conduct research analysis and storage of large health care databases for evaluating treatment utilization and outcomes in populations (ETrUOP).
The ETrUOP server is the key equipment to conduct research projects, which allows the efficient and collaborative work for a group of analysts and satisfies the data use agreements with data licensing agencies for data use.
HP DL380G7 server with 2 Intel Xeon X5675 CPUs (3.06 Ghz/6 cores/12Mb), 96 GB memory, 10 TB hard drive storage with fully encrypted double-tape backup and optical fiber channel connection. The server include Redundant Array of Independent Disks (RAID) level 50 hard disk configuration for data fault tolerance and performance. The RAID volumes and I/O configuration of the server is optimized for high SAS operation performance. The server has redundant power supplies and uninterrupted power supply. All unnecessary network ports and services will be disabled on the server for IT security. The research server is physically located in the state-of-the-art data center of the UNC Information Technology Service Research Computing, which is fully compliant with HIPAA and sensitive data research. The world-class central server room in the data center was designed to facilitate sensitive data computing with dedicated servers, including climate controlled air, gigabit network bandwidth, enhanced electrical power (220v), power conditioning (Uninterrupted Power Supplies), and restricted access controls.
Windows 2008R2 Server operating system with Terminal Services installed. Windows 2008R2 Server operating system will be routinely monitored and updated with patches, security updates and critical updates. Research software include SAS 9.20, STAT 10, ArcGIS 10, and Microsoft Office 2010 enterprise etc.
IT security policy and protocols
The dedicated server has (1) sufficient firewall and IP security, (2) a secure logon procedure, (3) security audit tracking, (4) local/group policy restrictions, (5) a strict password policy, (6) secure backups, (7) anti-virus protection, and (8) signed confidentiality statements by all users. This system enables approved researchers and designated collaborators to access and analyze confidential information from approved work locations. Network security and protection are top priorities. Software firewalls and Internet Protocol Security (IPsec) filtering have been installed to protect the server. All UDP and TCP network ports are permanently disabled, and access is restricted to a pre-determined set of individual static IP addresses and authorized office/lab desktop on UNC domain. All unnecessary networks and ports and services are disabled on the server. Network access to the server is limited to a Terminal Service connection. Terminal Services allows users to connect directly to the server from a remote location at which no data is transferred across the wire. The Terminal Service connection will be restricted in function through the use of local and group policies. Passwords will be encrypted. Full backups of the server are run every business nigh to double tapes in unbreakable 256-bit encryption. Anti-virus software has been installed to ensure data protection. Information Technology staff monitor event logs, including security audit logs, applications logs, and system logs. All staff with access to the server have HIPAA and IT security training for research data and read and sign a Confidentiality Statement.
Server administration and support
The server is administered and supported by the UNC ITS Research Computing. The UNC ITS Research Computing provides a world-class computing infrastructure as well as other technology tools and capabilities to support the research needs of UNC faculty and staff.
We have cross-disciplinary research resource files to support our advanced and cutting-edge research and pioneering analytical methods, including GIS files, US Census files, national health-care-provider file, ZIP code driving distance and time masterfile, drug database files, ICD-9 diagnostic code file, and health-care provider taxonomy codes database etc. These files have been standardized in SAS format and can be cross-linked to research datasets at ZIP code level or patient/provider level using our established SAS programs.
ETrUOP research resource files
|GIS boundary files||Boundary files for ZIP codes, MSAs, HSAs, HRRs, PCSAs, and States|
|GIS cross-walk files||ZIP codes to HRRs, HSAs, PCSAs, ZIP codes to States, Counties, MSAs, ZCTAs, States to Census Divisions and Regions|
|US Census files||2010 Summary Form 1 and ACS at census tract and block groups level||Geolytics Inc. for SF1|
|ZIP + 4||5 digit ZIP code Plus 4, centroids (Lat, Lon), 2010 census tracts and block||MelissaData|
|Lexicomp Integrated Drug Database||All pharmaceutical products in the US including NDCs, brand/generic name, active ingredients, strength, dosage form, therapeutic classification, manufacturer, OTC/Rx, generic, orange book classification, AWP etc.||Licensed by Lexicomp Inc.|
|ICD-9 database||All ICD-9 diagnosis codes and classifications.|
|National health-care provider file||NPPES national health-care provider master file. Include the name, NPI, address (including ZIP code) of practice or business, type, specialty etc of all organizational or individual health-care providers such as hospitals, pharmacies, SNFs, clinics, physicians, pharmacists, nurses etc registered with the Center for Medicare and Medicaid Services.|
|Health-care provider taxonomy codes database||NUCC Health-care provider taxomony codes database including the type, classification and specialty of health-care providers.|
|ZIP code driving distance and time (DTT) masterfile||The database contains over 34 million records of automobile driving distance, driving time, and average gas cost of unique routes between any 2 ZIP codes in the US.||Proprietary|
Gang Fang, Pharm.D., Ph.D., is an assistant professor in the Division of Pharmaceutical Outcomes and Policy. His research is centered in the areas of evaluating treatment utilization and outcomes in populations and pharmacoepidemiology especially in cardiovascular disease.