NDB Information…

Historical information. The National Databank for Rheumatic Disease (NDB) was founded in 1998 by Dr. Frederick Wolfe. Dr. Wolfe began computerized data collection in 1974 in his own rheumatology practice using programs he developed and wrote. From the start, every patient visit was entered into the computer in real time. In 1981, the clinic visit data were complemented by detailed survey questionnaires mailed to patients at 6 month intervals. This Databank, called the “Wichita Databank” is the largest and longest clinical rheumatology data base in existence any where in the world. The table below documents the number of patients and the number of visits and surveys contained in the Wichita Databank. The Wichita Databank has been the source of many important studies relating to the course and outcome of RA, OA and Fibromyalgia. In 1988, the Wichita Databank became a component Databank of the NDB.

Diagnostic Category

Patients

Visits and Surveys

Rheumatoid Arthritis

1,955

42,652

Osteoarthritis

3,602

24,315

Fibromyalgia

1,962

13,632

All Diagnoses

14,756

103,486


The genesis of the NDB. At the time of the founding of the NDB there were no true national, representative rheumatic disease Databanks. Beginning in the 1970’s, and with major expansion on the early 1980’s, the ARAMIS (Arthritis, Rheumatism, Aging Information System) Databank was the major RA Databank resource. Although state of the art in the early 80s, by the mid 1990s it had a number of deficiencies. There were only 4 centers collecting RA (2 at Stanford, 1 in Canada, and the Wichita Databank before dropping out of the program in 2001); the population had aged and had long duration RA and was not necessarily at risk to receive the newer drugs; and the number of patients available had shrunk and was under 2,000. In addition, there were no data using SF-36, WOMAC or psychosocial measures.

The NDB was founded with the idea of having a representative US sample that represented the practices of hundred of rheumatologists. SF-36 and other standard instruments were to be used; psychosocial questions added; and the Databank was to be expanded to approximately 10,000 patients such that the sample size would be adequate to test hypotheses at the individual drug level that reflected current treatment patterns.

Diagnosis

 

 

 

Group

Freq.

Percent

Cum.

 

 

 

 

RA

9806

76.95

76.95

OA

2502

19.63

96.58

Fibromyalgia 

436

3.42

100

 

 

 

 

Total

12744

100

 

In 1997, Dr. Wolfe made contact with US rheumatologists and asked them to enroll patients who they would see during a 30-day period. 13,486 patients were contributed, including all diagnoses. Beginning in 1998, individual rheumatology practices were incorporated into the Databank. Later, patients starting on leflunomide were added to the Databank at the time of the initial prescription. An inception cohort of RA patients enrolled as part of the ARAMIS project was also incorporated into the Databank. In 2000, patients completed detailed survey questionnaires: 9,806 RA patients, 2,502 OA patients, and 436 with Fibromyalgia completed detailed survey questionnaires. Throughout 2000, 2001, 2002, and 2003, additional patients have been added to the databank to expand its size and replace those lost by death and attrition.

In the Fall of 2003, Dr. Wolfe again contacted rheumatologists in the US and Puerto Rico and asked them to enroll patients who were seen during a consecutive 30-day period.

Staff. Dr. Wolfe serves as the overall director, but is assisted by an executive director, 2 programmers, a director of outcomes research, a project director, a Hispanic research director, and a team of outcomes staff and data verifiers. In all, the staff of the NDB totals 25 persons.

Data management. The primary method of conducting survey research is with a detailed 28-page questionnaire that is mailed to participants every January and July. When questionnaires are received they are scanned using Teleform software that has been specially programmed to address missing data and inappropriate responses. After data has been verified, it is transferred electronically to a SQL Databank. Depending on patient replies, additional follow-up is done. Hospitalization records and records of outpatient procedures are obtained. In addition, we verify reports of infection, cardiovascular disease, GI and liver disease, among other conditions. Validation of self-reported data is an important aspect of the NDB activities, and takes about 5 months to complete. The data are transposed to a convenient statistical package format for easy analyses, but are maintained in the SQL Databank as well. CD ROMs that reflect each 6-month mailing are produced for statistical analysis, and each CD ROM adds another 6 month’s data to the statistical analysis Databank.

Quality control. Several methods are employed. Questionnaire coding is checked by randomly selecting several questionnaires and hand checking the results in the SQL Databank with the individual questionnaires. Each month different section are evaluated such that in a 6 month cycle the entire questionnaire and coding is checked, validated and, where errors are found, corrected with implementation of new procedures and education. In addition, computer algorithms for range checking, date checking, inappropriate categorical sequence and missing data patterns are evaluated. Certifications scripts require all errors be corrected be for data is passed for general usage.

Facilities. The primary computing environment is based on the Microsoft SQL server that addresses over 100 GB of storage. Direct data collection via the Internet began in July 2001. The primary data entry method is by Teleform scanning software. All records are stored in an image Databank. We also store RA and OA x-rays in an image Databank. These images are acquired using high definition Lumysis radiographic scanners. A serum and whole blood Databank is available and fully operational. Data are backed up to tape daily and stored off site.

Content. The Databank collects data on current clinical status, including HAQ, ClinHAQ, SF-36, WOMAC, pain, fatigue, global severity, anxiety, depression, quality of life, self reported joint counts, among other items. All drug usage is documented and side effects are reported and validated. A full range of utilization variables is available. Co morbidity and specific questions regarding infection, GI disease and cardiovascular disease are investigated in detail. Symptoms and demographic data, including work and indirect costs are part of the content.

Purpose. The purpose of the NDB is to advance knowledge regarding the outcomes of rheumatic diseases, including predictors of outcome and factors related to treatment, treatment response and treatment adverse effects. The NDB is now collecting its 13th cycle of data (6.5 years). A number of manuscripts are already in press, and preliminary abstracts have been published in 1998, 1999, 2000, 2001, 2002 and 2003 in Arthritis and Rheumatism. In 2000, the NDB published 25 abstracts, more than any other US center.

For additional information contact Dr. Fred Wolfe (fwolfe@arthritis-research.org) or the executive director, Rebecca Schumacher (rebecca@arthritis-research.org).