Interview with Dr A.J. Collins

Dr A. J. Collins Nephrology Analitical Service University of Minnesota Minneapolis, USA

Zoccali: What lead you into be interested in epidemiology and biostatistics? Was this an early choice in your career?

Collins: As a clinical nephrologist and applied math major as my area of interest in Medicine in my undergraduate education, I chose artificial organs research and applications. As I did clinical research in dialysis therapy, long-term outcomes of the patients became an important issue. My mentor, Dr Fred Shapiro, who had developed a large dialysis operation in the upper Midwest of the United States, stressed that we should assess the impact of our interventions and treatments on the ultimate survival of the patients. Therefore, I grew up in a training program that emphasized a careful analysis of clinical data to determine the right course of action. I ultimately took over the Regional Kidney Disease Program 6,000 patient detailed dataset that was initiated in 1975 before databases were considered important. This ultimately led to the development of a large analytical group at the Minneapolis Medical Research Foundation at Hennepin County Medical Center, University of Minnesota with the subsequent technical expertise being developed to handle the United States federal data.

Zoccali: Who was the colleague/teacher that had a major influence on your professional and career choices?

Collins: There are four mentors and colleagues that had a major influence on my professional development and career. Dr Fred Shapiro, Emeritus Professor of Medicine at the University of Minnesota and founder of the Regional Kidney Disease Program in 1966 at Hennepin County Medical Center, recruited me into nephrology and was a key factor in my subsequent interest in artificial organ research and database analysis. The second was Dr Prakash Keshaviah, an outstanding bio-engineer and research colleague who I worked with for over 18 years. Dr Keshaviah’s systematic approach to the analysis and design of clinical trials in renal replacement therapy, as well as his expertise dialysis were instrumental in my career. The third was Dr Eli Friedman from Downstate Medical Center in New York who both challenged and guided me in many areas of my career. The fourth was Dr Christopher Blagg in Seattle with the Northwest Kidney Center, who for many years was considered our sister organization because of the collaboration between Minneapolis and Seattle on many areas of dialysis treatment.

Zoccali: Could you briefly tell us how the USRDS is organized and which is the backbone of the USRDS managerial structure?

Collins: The USRDS is a contract with the National Institutes of Health (NIH), National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). The two Project Officers at the NIH that oversee the USRDS are Drs Paul Eggers and Lawrence Agodoa. The USRDS operates under an interagency agreement between the NIH and the Centers for Medicare and Medicaid Services (CMS). Almost all the data that is used by the USRDS comes from CMS along with information from the Centers for Disease Control and Prevention. The managerial structure of the USRDS is laid out in the Introduction section of the Annual Data Report Atlas. The major operational components are Deputy Director Bertram Kasiske, MD, Director of Information Services, Data Systems and Quality is Shu Chen, MS, Editorial Production of the Annual Data Report is Susan Everson, PhD, Business Operations is Kim Miller and the Co-Investigators. The Co-Investigators are broken down into areas of specialty such as Epidemiology, Medical, Economic, and the Biostatistics group. The Biostatisticians have a central role in all analyses related to the quality, design, and methods used. This group is headed by Thomas Louis, PhD at Johns Hopkins School of Public Health. The other Biostatistical Co-Investigators have extensive expertise in theoretical biostatistics, as well as Bayesian analysis. There are four Special Studies Centers (SSC) affiliated with the USRDS. The first awarded was the Cardiovascular SSC headed by Charles Herzog, MD a cardiologist at Hennepin County Medical Center. The Economic SSC is headed by Dr Lawrence Hunsicker at the University of Iowa. The Nutrition SSC is located at the University of California, San Fransisco and directed by Dr Glenn Chertow. Dr Nancy Kutner heads the Quality of Life/Rehabilitation SSC at Emory University.

Zoccali: The USRDS is a complex structure. How data quality control is ensured in the USRDS?

Collins: Data quality is a central issue for the USRDS to address. Certain parts of the U.S. federal data are extremely accurate and other areas are more challenging. Since 1995, all individuals that initiate end-stage renal disease (ESRD) treatment are required to report a Medical Evidence Form to CMS. Each of these forms contains the patient’s date of birth, gender, race, primary cause of renal failure and various lab values. These are among the most complete patient information within the dataset. The first service date for ESRD treatment is central to all outcome analyses and tracking systems. Other aspects of the information on the Medical Evidence Form, however, are less than complete. For instance the comorbidities that are listed on the form account for only one-third to one-half of all the comorbidities the patients carry as they enter ESRD treatment. Biochemical data is available on approximately 75-95% of the patients. This biochemical data at initiation of treatment consists of serum creatinine, BUN, hemoglobin, hematocrit, and serum albumin. Height and weight data are also fairly complete on this form.

The next largest area is the modalities of treatment. Since Medicare is the primary, but not the only, payer for ESRD services in the United States, there are a number of challenges within these patient populations to determine their treatment modality. For those individuals that are over 65 years of age, there is fairly complete information on the modalities of treatment over time. However, we are unable to determine the exact modality of most patients less than 65 years old in the first 90 days because of a lack of availability of information from the Medical Evidence Form until after the first 90 days of treatment. The patient population where it is very difficult to determine the modalities of treatment is the managed care patients and those individuals covered by private insurance. This accounts for 30% of the patients in the incident population and 22% of the prevalent patients. At December 31st of each year, we have a fairly accurate count of the patients by modality in this group but not the changes that occur within the year. This gap in information is being rectified now for next year’s report.

In order to fill those gaps, special data collection studies are done by CMS and by USRDS. Other data that is heavily utilized by the USRDS are events such as heart attacks, causes of hospitalization, procedures, and revascularization procedures. Procedures are very well documented in the federal data because of the requirements of accuracy in billing and the severe penalties for fraud and billing abuse. Death information overall is reported fairly completely, however, cause specific deaths are sometimes difficult to ascertain. Almost 20% of the primary causes of death may be for miscellaneous causes or unknown because they occur outside the hospital setting. Overall, the number of individuals treated, the types of modalities, age, gender, race, and primary cause of renal failure are accurate. Hospitalizations and cause-specific hospitalizations are fairly complete as well as the major procedures in the Medicare population.

Zoccali: You have made a tremendous effort to produce clinical performance measures. Knowledge of these data has a great potential for improving the management of patients with ESRDS. Is there any particular weakness or any problematic aspect that you would like to solve very soon on this matter?

Collins: Clinical indicators of care cover a broad aspect of the USRDS Annual Data Report. We have currently focused on the measures that are typically included in the National Kidney Foundations K/DOQI guidelines, as well as standard preventive health care measures that are recommended for the general population including diabetes care, influenza vaccinations, Hepatitis B vaccinations, and lipid monitoring in patients with cardiovascular disease. Certain parts of these measures are extremely accurate and some may be underreported. In the United States for example, free vaccinations for influenza may occur or patients may receive vaccinations outside the Medicare system. Since hemodialysis patients are seen three times a week, there is a considerable incentive to provide these services since they are readily available and providers can bill for these services. Hematocrit data is available extensively on the hemodialysis population but only half of the peritoneal dialysis patients receive erythropoietin, which requires the reporting of a hemoglobin/hematocrit on the claims for this service. Other data that would reflect the erythropoietin and non-erythropoietin treated peritoneal dialysis population comes from CMS Clinical Performance Measures Project. To the extent that certain preventive health care measures may be underreported within the Medicare system, primary data collection from a random sample of individuals in the CMS Clinical Performance Measures Project would more accurately reflect the true utilization and results.

Zoccali: Is there any additional Clinical Performance which you feel important adding to those already implemented?

Collins: The clinical care of ESRD patients, as well as chronic kidney disease patients, is an open area of investigation for the USRDS. We are currently evaluating preventive health care measures in the general Medicare chronic kidney disease population, as well as the ESRD patients. The care and complications of transplant patients are also an important area of which our Deputy Director Dr. Kasiske is investigating. Peripheral vascular disease and amputations, particularly in the diabetic population and those with advanced vascular disease, are important areas of morbidity. Timeliness of evaluations and the types of procedures before amputations occur are an important area to consider and the USRDS is attempting to define methods to assess this aspect of patient care.

Zoccali: Which developments have you planned for the USRDS for the next few years?

Collins: The USRDS designs the Annual Data Report every year and determines the data analysis for presentation. The appropriate text is then added to complete the report. The USRDS does not pre-advertise the types of analyses it will do but works with the Project Officers and External Advisory Committee on areas of investigation. Overall, the USRDS will look at the quality of care that the chronic kidney disease and ESRD populations receive.

Zoccali: Is epidemiology of renal diseases and related complications formally included in training programs of renal physicians in the USA? Do you feel that more emphasis should be given to this discipline?

Collins: The epidemiology of ESRD is only rudimentarily covered in post-graduate training. These training programs should familiarize physicians in general and nephrologists in particular on the use of epidemiological data since considerable information is being published in this area and they should be familiar with these analytical tools.

Zoccali: Have you taken any initiative for attracting the interest of nephrology trainees and young nephrologists in general on the USRDS and on Registry Research?

Collins: The USRDS receives requests for assistance in study design from a number of investigators, including nephrology trainees. Specifically, the University of Minnesota and Hennepin County Medical Center are involved in a collaborative Clinical Science Masters degree program within the School of Public Health. Investigators at the USRDS are part of the training curriculum and on the faculty at the School of Medicine. Many of these trainees do projects with the federal data.

Zoccali: Have you planned collaborative (comparative) studies with other dialysis registries?

Collins: A number of registries around the world have been discussing studies, particularly in the area of diabetes and the quality of care of the diabetic population worldwide. Several registries with the USRDS have collaborated on cancer in the dialysis patients historically.