Introduction

What follows are reports of New York State hospital and physician specific volumes for a number of surgical and diagnostic procedures in 2002 (the most recent year for which data is available). These reports are intended to help consumers make more informed decisions when they select their health care providers. Consumers should ask a physician how many procedures she or he has performed in the last year and compare that number to these reports. If there is a large difference in either direction ask the physician to explain why.

Information about a procedure performed outside of licensed hospitals or ambulatory surgery facilities, for example in a physician's office, is not required to be reported to the state. For some procedures, such as colonoscopy, an increasing number are done in facilities, such as a physician's office, that are not required to report. Therefore, a low volume may not be an indication of lack of experience, but rather the lack of reported information. We have included a cautionary note to this effect in each report introduction where appropriate.

Except for coronary artery bypass surgery and more recently, coronary angioplasty, New Yorkers have not been provided with information about the performance characteristics of a hospital or doctor by the state. Hospital and physician performance characteristics include, among other things, how often a hospital or physician performs a procedure and/or treats a specific medical condition. New York is one of a very few states where any performance information is made publicly available.

The Center for Medical Consumers is committed to unrestricted public access to all information about health care providers collected by government with public funds for the purpose of assuring safety and quality of medical care. However, we recognize that such information can be more useful to consumers when presented in a context that helps them better understand its' strengths and limitations. For that reason, we strongly suggest that you read the remainder of this narrative prior to referring to the hospital and physician-specific volume information that follows.

Volume or Experience and Quality

The relationship between volume and experience and quality has not been firmly established for every procedure or diagnosis. But a growing number of published studies do provide evidence that volume and high quality outcomes are related for a number of diverse medical conditions and interventions such as cardiac surgery, cancer and AIDS treatment. In light of such studies, and because it makes common sense, the Center believes it is reasonable to view procedural and diagnostic volume as an important indicator of a hospital's or physician's experience. However, consumers should be aware that a high volume of surgical procedures does not guarantee a high quality outcome.

Because comparative hospital and physician volume information does not require the risk adjustment that is necessary when reporting comparative outcome information, the data collection, analysis and publication of volume data are far less resource intensive than what is required for reporting medical outcomes.

Outcome and Quality

There has been a great deal of interest expressed by consumers, purchasers and government policy makers in comparative medical outcome reporting. However, the process of developing valid risk adjustment methodologies and obtaining a consensus concerning their validity has been contentious. As a result, progress in developing risk adjusted comparative performance and outcome reports has proceeded at a snail's pace nationally. Besides New York, Pennsylvania, Virginia and California, few other states appear to have any interest in producing hospital and/or physician-specific outcome information. The few states cited above have only produced a handful of reports and have had to devote significant resources to do so. Additionally, in an era of downsized government, state health departments in general may lack the resources necessary to either begin or to expand outcomes reporting.

Why These Reports Do Not Rank Hospitals or Physicians

We chose not to list hospitals or physicians in rank order by volume. Instead they are listed alphabetically by name or by location. A ranking by volume would suggest that the first hospital or physician listed (and the one with highest volume) is the best, and the second listed is the second best, and so on. Even if the relationship between volume and a good outcome is strong for a given procedure there are likely an optimum number of procedures beyond which there is no or very little effect on quality. For example, experience shows that hospitals should perform a minimum of 200 cardiac bypass surgeries annually to achieve good outcomes. But, we have no indication that a hospital doing more than the threshold of 200 surgeries will, as a result, achieve any better outcome. There is, we believe, an additional reason not to rank order physicians by volume. A very high volume could be an indication that a physician overuses a procedure. Overuse, the provision of unnecessary medical care, is poor quality medicine.

Very Low Volume Physicians

One of our overall findings was that a substantial percentage of many procedures are performed by individual physicians at very low volume, often only one or two annually. Because we generally selected procedures with fairly large overall volumes (both numbers of procedures and numbers of physicians performing them) for our initial reports, we decided that the reports would be less "user-friendly" to download if every physician doing the procedure in 2002 was listed. However, we also believe that consumers should be able to access all the information available. So we created two report formats for physician-specific volume information: a "user friendly" one with a minimum volume cutoff and one listing all physicians who performed one or more procedures in 2002.

For the "user-friendly" report formats when statistically appropriate, the volume criterion for physician inclusion is set at a minimum of 30 procedures annually — a threshold that is used, for example, in New York's physician-specific cardiac surgery outcome reporting. In some cases a lower threshold was established because the statewide volume of the procedure was low. Since there are approximately 260 licensed acute care hospitals in New York State we saw no advantage to creating two formats for hospital volume. As a result, each hospital-specific volume report lists every hospital filing a SPARCS report with the relevant ICD-9M code. Conversely, if a hospital is not listed, it did not report any activity for that specific ICD-9M code in 2002.

Sources of the Data

Our reports of 2002 hospital and physician-specific volume for selected procedures are derived from the New York State SPARCS database and ambulatory surgery database, which comprise information filed by hospitals and licensed ambulatory surgery facilities with the New York State Health Department for every discharged patient. The New York State Health Department maintains the SPARCS and ambulatory surgery data bases and annually aggregates these reports from licensed health care facilities into so-called "administrative data files" which are available for purchase. The administrative data files contain no individually identifiable patient information.

We sorted all observations where the indicated ICD-9 code appeared either in the primary procedure code field or any of the fourteen "other" procedure codes fields. The introductory narrative for each report describes how the sort was done.

In order to prepare the reports of physician-specific volume, the Center obtained a computer file from the State Education Department's Office of the Professions (OP) that matched physician license numbers with physician names. Both the SPARCS and ambulatory surgery databases use physician license number as the identifier. Our reports list volume information only for those physicians for whom a license number was matched with a name. It is our conclusion, after some investigation, that the absence of a physician name to match a license number that is active in the 2002 SPARCS or ambulatory surgery administrative file is most likely an indication of one of the following: (1) matching problems that may arise because of the two different databases and agencies; or (2) the physician did not renew his or her registration for 2002; or (3) the physician has elected inactive status for 2002; or (4) the physician had her or his license revoked or suspended or otherwise sanctioned in 2002.

Each report includes a statistical analysis that includes bar graphs of hospital and physician-specific volume as well as other statistical information, such as percentiles and means, to facilitate comparing the volume experience of a hospital or physician to all hospitals and physicians in the state performing the procedure.

Accuracy of Reports

The accuracy of data collected through the state-mandated SPARCS system has been the subject of much discussion over the years. There is some evidence that the accuracy of the SPARCS database has been steadily improving over the years. In addition, every reporting facility is given the opportunity to review and correct information in the database. However, the state ultimately is dependent on licensed health care facilities to provide complete and accurate information. The states own resource constraints limit what can be done to validate the reported data. And, unfortunately, changes in New York State hospital reimbursement methodology over the past few years may erode the incentive for providers to rigorously comply with SPARCS reporting requirements.

The same concern exists about the accuracy of the ambulatory surgery database. One difference is that it is less mature than SPARCS and its accuracy has been subject to less review over the years. The Department of Health agrees that there are improvements that could and should be made to assure better quality of data for both reporting systems.

The Center for Medical Consumers must rely on the data provided by the state and cannot independently assess or improve its accuracy. However, we take some comfort in the fact that SPARCS data is regularly used by the state and often purchased by independent researchers and industry for analysis in pursuit of a variety of academic and commercial objectives.

We feel confident that our report accurately reflects the information about volume currently available from New York State databases. Hopefully, as the public gains more routine access to information about comparative performance characteristics, licensed health care facilities and professionals will hopefully be motivated to do what is necessary to assure that the information they report to the state is of the highest possible quality.

A Caution About Physician Procedure Volume Reporting

There is an important cautionary note about these physician volume reports. The physician who actually performed the procedure is supposed to be the one whose license number appears in the operating physician field on the report a hospital submits to SPARCS or the ambulatory surgery database. However, it is well understood that this is not always the case - for example, a chief of service may sign off in the medical record on procedures actually performed by other physicians and as a result it is his or her license number that appears on the submission.

The practice described above is not in accord with reporting criteria established by the Department of Health. But, unfortunately there is no practical way for the state to routinely evaluate the extent of this problem or to enforce proper reporting. Such inaccuracies in reporting may result in the volume information in our reports either overestimating or underestimating the actual volume of an individual physician. Our advice to consumers is to ask a physician how many procedures she or he has performed in the last year and then compare that number to these reports. And, as we pointed out earlier, an increasing number of procedures are being performed in doctors' offices, which are not licensed by the state and are therefore not subject to reporting requirements of any kind.

Research Details

A commercially available, command driven, statistical software program known as STATA (Version 8 SE) was used to manage, analyze and graph the data obtained from the Department of Health and produce a detailed statistical analysis of the per hospital and per physician volumes. For more information about the software package visit www.stata.com The International Classification of Diseases 9th Revision, Clinical Modification 5th Edition (ICD-9-CM) was used to identify procedure codes reported under SPARCS.

The Center's research consultant, Enrique Geiger, sorted and analyzed the original administrative file of all hospital and ambulatory surgery discharges from licensed facilities in 2002, ran numerous tests to determine the accuracy of our sorting and aggregation by ICD-9-CM codes and New York State hospital identifiers and physician license numbers. Mr. Geiger erred on the side of caution in ensuring that our extraction of the selected information contained in these reports was accurate. We also had informal consultations with health services researchers about the data, its limitations, our analysis and presentation. Arthur A. Levin, MPH, director of the Center for Medical Consumers, determined which procedures and ICD-9-CM codes to include and is responsible for this introductory narrative.

Notice

The Center for Medical Consumers does not make referrals or recommendations to specific doctors or hospitals and these reports are not intended to recommend any particular doctor or hospital. The information contained in these reports is derived solely from reports made by hospitals and ambulatory surgery facilities licensed under Article 28 of the New York State Public Health Law to the Department of Health.

© 2005 Center for Medical Consumers
The Center for Medical Consumers cannot respond to inquiries regarding individual health concerns.
Our Web site is updated during the first two weeks of each month.
Please send any comments or questions to medconsumers@earthlink.net