| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Vascular and Interventional Radiology |
1 From the Department of Radiology, Massachusetts General Hospital, White 270, 55 Fruit St, Boston, MA 02114. From the 2000 RSNA scientific assembly. Received November 21, 2000; revision requested January 2, 2001; revision received March 5; accepted April 6. Address correspondence to P.F.H. (e-mail: phahn@partners.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Date, organ location, and type of all 21,324 procedures performed between October 1990 and September 2000 were recorded in a database; also included were patient age and inpatient or outpatient status. Because genitourinary procedures were not included during the first 4 years, nephrostomies were added retrospectively. Yearly interventional caseload was compared with the department caseload and the assignment of new medical record numbers. Trends in individual procedure location and type were analyzed, as well as patient age, inpatient status, daily caseloads, and day of the week when the procedure was performed.
RESULTS: Caseloads have increased 10.8% per year, exceeding increases in radiology department and hospital activity. Abdomen, outside of a specific organ, was the most common location; catheter deployment was the most common procedure. Abscess drainage, placement of chest tubes, and nephrostomies have increased, but biliary drainages have decreased. Inpatients outnumbered outpatients in all years except 1995, but the trend was toward an increase in the proportion of outpatients. The average patient was 59.6 years old, with average age diminishing. Friday was the busiest day, but weekend procedures have increased.
CONCLUSION: Nonvascular interventional procedures have increased, with more currently being performed on weekends.
Index terms: Interventional procedures • Radiology and radiologists, departmental management
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The database has been maintained continuously for 10 years, beginning prior to October 1, 1990, and it is ongoing as of September 30, 2000. Because fiscal years (FYs) begin on October 1, we describe the database in terms of FYs 1991 through 2000.
The structure of the interventional database and the method of data entry have been previously described (1). In the database, the patient name, medical record number (MRN), date of service, and two descriptors of each procedure, its location and type, were recorded. After the first 2 years, the name(s) of the operators were also recorded. Data were entered daily from a blackboard that was used to organize the workflow and then were corrected each evening as part of sign-out rounds. The following is an example of data entry: Two catheter drainages of abdominal collections performed in the same patient on the same day would be recorded as a single procedure; if a diagnostic thoracentesis were also performed, that would be entered as a second procedure. Procedures such as renal biopsies in which a nonradiologist obtained the samples while a radiologist provided imaging guidance were not included in the database. The imaging modality used for guidance of each procedure was also not recorded.
Preparation of the database containing interventional procedures proceeded prospectively with the following exceptions. When the system operator (P.F.H.) returned from any absences, he checked reports and resource schedules in the radiology information system (IDXrad; IDX Systems, Burlington, Vt) against the information entered in the database during his absence. Necessary corrections or additions were made at that time. The age and the inpatient or outpatient status were not prospectively recorded in the database. These data were entered retrospectively by means of cross-matching interventional entries in the database with those in the radiology information system listings according to MRN and date of service. The cross-matching process also served as a check on the accuracy of database entries.
First pass accuracy of unit numbers entered in the database ranged from 1,283 of 1,360 (94.3%) in 1991 to 2,545 of 2,560 (99.4%) in 1998. Overall, the MRN and date of 551 of the 21,341 procedures (2.58%) originally recorded in the database failed to match with an entry in the radiology information system. These cases were individually reviewed with hospital records. Most cases turned out to have MRNs incorrectly recorded by one digit or by a simple transposition of two digits. A total of 17 cases were expunged from the database as a result of this verification process.
A final exception to the prospective nature of data acquisition concerns genitourinary (GU) procedures. Until 1995, there were separate divisions of gastrointestinal and GU radiology. Procedures performed by the staff in the GU division were not recorded in the database maintained by the gastrointestinal division. Consequently, there are few database entries for renal-adrenal procedures performed before 1995. In fact, the paucity of GU procedures performed when we started the database motivated us to combine the adrenal and renal cases into a single category. Bladder and prostate categories had to be added to the database in 1995, when data for these locations began to be acquired.
We took two steps to mitigate the deficiency of GU procedures. First, some of the key analyses were performed twice, both including and excluding the GU procedures. Second, we used hospital sources to enumerate all of the nephrostomies and renal abscess drainages performed prior to 1995. These efforts included searching texts of radiology reports and printouts of resource schedules, working backward in some cases from subsequent tube manipulations to identify earlier nephrostomy insertions. Complete retrospective acquisition of all lost GU procedure data, such as biopsies, suprapubic cystostomies, and tube manipulations could not be accomplished. The retrospective data for nephrostomies and renal abscess drainages obtained prior to 1995 were not added to the database but were maintained for a separate analysis of decade-long GU trends.
Data Analysis
For the 10 years, we determined the rate of case accrual in the database over time. This measure of yearly interventional radiology activity was then related to measures of departmental and hospital-wide activity, as described later. We also determined the distribution of cases, both as to the location of each procedure (organ or body area) and the type of procedure (eg, biopsy, catheter placement). Trends within these categories during the 10 years were also sought. For each of the 10 FY, we computed the number of inpatient and outpatient procedures and compared the ratio as a function of time. The age profile of the patients was also determined. Additional information extracted from the database concerned patients at the extremes of procedure volume (one procedure or 
20 procedures), the number and experience level of operators who performed the procedures, and the distribution of procedures on weekdays and weekends. No attempt was made to assess procedure outcome.
Two measures of departmental and hospital activity were compared with the interventional activity recorded in the database. The measure of radiology department activity was the volume of cases accessioned in the radiology information system. Assignment of new MRNs served as a relative measure of hospital activity. The rate of new MRN assignments was selected because it is a readily available statistic that reflects both inpatient and outpatient services across the entire enterprise, and MRN assignment has been a consistent process for decades. The largest MRN that appeared in the radiology information system each September was used to determine the number of MRNs assigned during that year by subtracting the largest MRN from the preceding September. Data for both measures were derived from the radiology information system listings, which were also used, as described previously, for the determination of patient age and inpatient or outpatient status.
Statistical Analysis
Trends in the rate of case accrual were modeled by means of linear regression. The increase or decrease was considered significant if the 95% CI for the slope of the regression line excluded zero. The one-sided z test for binomial proportion was used to test differences in the number of cases that occurred on various days of the week.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Increase in Database Volume
Each year, except for one, the number of entries in the database increased compared with the previous year. Twenty-three fewer procedures were performed in 1997 than in 1996, a decrease of 1.0% (23 of 2,217 cases). On average, 230.5 more procedures were performed during each succeeding year (95% CI: 176, 285; P < .001), an average of 10.8% (230.5 of 2,132 procedures). Overall radiology department and hospital activity also increased during this period. However, the increase in interventional procedures exceeded both of these other measures. In the 10 years of the study, we performed six special procedures for every 1,000 studies completed by the department as a whole (0.6%). This ratio increased by an average of 0.051% each year (95% CI: 0.041%, 0.061%; P < .001). We performed 31 special procedures for every 1,000 newly issued MRNs (3.1%). This ratio increased by an average of nearly 0.15% each year (95% CI: 0.097%, 0.196%; P < .001). Thus, growth in interventional procedures exceeded growth in both radiology department and hospital activity. Absolute and relative increases in interventional activity are displayed in Figures 1 and 2.
|
|
|
|
|
|
Indwelling catheter deployment (5,193 cases) was defined as placement of a catheter that was left in the patient at the conclusion of the procedure. This category excludes transient deployments such as therapeutic thoracentesis or paracentesis. When the abdomen category was taken to include the pelvis, mesentery, and retroperitoneum, the abdomen was the most common location for indwelling catheter deployment. Abdominal catheter deployments, accounting for 1,959 procedures, were mostly performed for abscess drainage. Abdominal catheter placements increased at a rate of 12.6 cases per year (6.4% per year; 95% CI: 8.5, 16.7; P < .001). Chest tubes, including a few mediastinal tubes, composed the next most common type of catheter deployment (1,359 cases). Placement of chest tubes increased rapidly during the 10 years, at a rate of 15.4 cases per year (11.3% per year; 95% CI: 11.3, 19.5; P < .001). Biliary drainage (499 cases) was the third most common procedure in which a catheter was deployed. Unlike abscess drainage and chest tubes, biliary drainages decreased during the 10 years, averaging nearly 2.6 fewer cases each year (5.2%, 95% CI: 0.4, 4.7; P < .03). Had nephrostomies been entered into the database from the beginning, the number of nephrostomies would have exceeded biliary drainages (see next section).
GU and Non-GU Procedures
Of 21,324 procedures performed in the 10 years, 2,932 were specifically GU, comprising renal-adrenal, bladder, and prostate locations. Thus, the remaining 18,392 procedures may be considered non-GU. Because a separate division of our department performed GU procedures until 1995, we analyzed the non-GU procedures separately to determine whether the growth in procedures could have been due to the GU procedures added at that time. The absolute growth and the growth relative to the radiology information system and MRNs increased at 145.6 cases per year, 0.029% per year, and 0.055% per year, respectively. All of these positive rates are statistically significant (P < .05).
Using retrospective data for procedures performed prior to 1995, we identified 905 procedures in which a catheter was left in the kidney. These were mostly nephrostomies, but included without distinction in the database a few renal abscess drainages. We compared renal catheter drainages with non-GU procedures because these were not affected by the inconsistent recording of GU procedures during the 10 years of the study. The percentage of renal catheter drainages was 4.9% (905 of 18,392) of the non-GU procedures. Compared with non-GU procedures, the minimum yearly number of renal catheter drainages was 2.9% (48 of 1,662) in FY 1995, and the maximum was 7.0% (141 of 2,006) 3 years later, with no identifiable linear trend (P = .39). However, in absolute terms, renal catheter drainages increased at an average rate of almost 10.2 cases per year (11.2% per year; 95% CI: 3.7, 16.6; P < .01).
Patient Age
The age of all patients in the database is known. The average age of patients at the time of their procedures was 59.6 years ± 19.0; 4,939 (23%) of 21,324 procedures were performed in patients in the 7th decade of life, the most common decade for procedures. Thirty-four procedures were performed in 19 patients who were younger than 1 year of age. Three patients underwent four procedures as centenarians (oldest, 105 years). There was a tendency for the roughly bell-shaped curve of frequencies to move to the left during the years in which data were obtained, indicating that we have seen more patients who were younger. Indeed, the linear regression line for average age diminishes with a slope of -0.238 year for each FY (95% CI: -0.038, -0.438; P = .001). SD ranged from 17.74 to 20.10, without an identifiable trend over time.
Inpatient or Outpatient Status
For nine of the 10 years, inpatients outnumbered outpatients. Overall, 12,478 procedures were performed in inpatients and 8,846 in outpatients, a ratio of 59% to 41%. Only in FY 1995 were there more outpatients (1,047 or 56%) than inpatients (834 or 44%). The number of procedures performed in inpatients reached a nadir anomalously in all of our radiology department in 1995, when 73,138 (20%) of the 365,895 radiology information system accessions were inpatient studies versus 34% (1,096,369 of 3,216,437 studies), which was the average for the other years. However, the trend was for the inpatient procedures to decrease relative to the outpatient procedures, so that in FY 2000 the numbers of inpatient and outpatient procedures were almost equal (1,711 vs 1,723) (Fig 4). This trend is statistically significant (P < .05).
|
Twenty-five patients underwent 20 or more procedures. The greatest number of procedures for any one patient was 51, followed by 45. Together these 25 patients underwent 660 procedures. The most common procedure performed in this 25-patient cohort involved nephrostomy tube manipulation (160 procedures), followed by therapeutic paracentesis (101 procedures).
Operators Performing Interventional Procedures
Beginning with FY 1993, the names of the operators who performed each procedure were recorded in the database. These individuals included residents, fellows, and staff members in the department of radiology, as well as one nurse-practitioner. Between October 1992 and the end of September 2000, there were 140 different individuals who participated in nonvascular interventional procedures that were recorded in the database. Four staff radiologists each had responsibility for more than 2,000 cases. Of the 18,525 procedures performed during the 8 years from October 1992, these four radiologists participated in 10,104 (55%) of them. Four junior members of the staff each had been involved with between 1,000 and 2,000 cases; many of these occurred during their training.
Weekday and Weekend Procedures
Of the 21,324 procedures performed in 10 years, 1,207 (5.7%) were performed on Saturday or Sunday, and 20,117 were performed on weekdays. On average, 7.7 procedures were performed each weekday, increasing at a rate of 0.8 cases per weekday (10.4%; 95% CI: 0.6, 1.0; P = .001). Among weekdays, the busiest day was Friday, when 4,218 (21.0%) of the 20,117 weekday procedures occurred. Tuesday, Wednesday, and Thursday each trailed Friday by less than 1.0%, a difference that barely reached statistical significance (z = 1.78; P = .038). A total of 3,640 (18.1%) of the 20,117 weekday procedures were performed on Monday, less than on the other days (z = -6.72; P < .002) because of the preponderance of Monday holidays.
The tendency for procedures to be performed on Fridays was particularly pronounced during the first 2 years. During FYs 1991 and 1992, Fridays accounted for 23.8% (312 of 1,309) and 23.3% (325 of 1,393), respectively, of the weekday procedures. This excess volume of procedures performed on Friday versus Tuesday, Wednesday, and Thursday was statistically significant (z = 4.22; P < .002). In 1992, we noted the increased procedure volume on Fridays and moderated its effect by more carefully scheduling elective cases.
In the first 3 years, the percentage of procedures performed on weekends did not exceed 4.5%. However, for FY 1994 and after, 6% or more (maximum, 216 of 3,144; 6.9% in FY 1999) of the procedures were performed on weekends, until FY 2000 (161 of 3,433; 4.7%). The incidence of weekend procedures was reanalyzed for non-GU cases. A total of 1,070 (5.8%) of the 18,392 non-GU procedures were performed on weekends. Non-GU procedures performed on weekends increased from less than 4.5% in all years through FY 1993 to more than 6% (maximum, 145 of 2,006; 7.2%) in all years after that until FY 2000 (140 of 2,768; 5.1%). Thus, the merging of GU procedures into the database does not account for the increase in weekend procedures seen during FY 1994.
There were 11 days when 20 or more procedures were recorded. Three were in FY 1999 and the rest in FY 2000. The maximum number of procedures performed in one day was 25 and involved 20 patients and six staff radiologists.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Many radiology departments are set up differently from ours (2,3). For example, in many departments, a single vascular and interventional radiology section performs all of the nonneuroradiologic interventions as part of a day that also includes vascular cases. In other departments, procedure responsibilities break along modality lines, with ultrasonographers performing biopsies and drainages with US guidance, while computed tomography (CT) cases are performed by a section also responsible for diagnostic CT. This divergence reflects the diversity of skills required to perform these procedures, from Seldinger guide-wire techniques to CT-guided placement of radio-frequency electrodes.
The skills and knowledge unique to nonvascular intervention have been recognized in that one-third of the three-part examination for the certificate of added qualifications in vascular and interventional radiology tests the candidate’s preparation for nonvascular procedures (4). Although our franchise excludes breast, lung, and bone biopsies, as well as joint aspirations and injections, our data provide an extensive compilation of the volume and composition of nonvascular interventional radiology procedures in a large teaching hospital. Moreover, the database embodies the changes that have occurred within that caseload volume and composition during 10 years that saw major changes in American medicine, particularly in radiology.
Our hospital underwent a transition toward managed care during FYs 1994–1995. There was a period before the conclusion of appropriate managed care contracts when the hospital inpatient census decreased. This is reflected in the anomalous preponderance of outpatients in the database for FY 1995. Overall, radiology department revenue diminished, and staffing levels were trimmed. This period coincided with a bleak outlook in job prospects for radiologists (5). The volume of procedures flattened and then actually diminished slightly in FY 1997, before resuming its upward climb once staffing levels had recovered (6). Because of the falling volume in GU among other sections, the department undertook a consolidation that resulted in a single division charged with all abdominal imaging and intervention. Emphasis on a shortened length of stay meant procedures could no longer be deferred until regular business hours; therefore weekend procedure volume increased (7). As the recovery proceeded, there was a shift toward performing procedures in outpatients whenever possible.
Job-related psychologic stress, especially related to work overload, has recently been shown to be common among radiologists (8). Several features of our database analysis show the potential for the radiologists to experience stress. The most obvious stressor is the sheer increase in the number of cases, which more than doubled during the 10 years of data collection, increasing at an average rate of nearly 11% per year. Moreover, there has been an increase in off-hour procedures because of the pressure to expedite patient discharge.
Another potential for stress is the increasing proportion of outpatients undergoing interventional procedures, a trend noted in radiology as a whole (9). A shift toward outpatient procedures might mean a reduction in the average effort for a given patient interaction. Indeed, outpatient procedures tend to be elective, and outpatients are more robust and in more stable condition than are inpatients. However, outpatient procedures carry their own stresses. These include timing the patients’ arrival, ensuring proper patient preparation (eg, fasting and adequate blood work), obtaining the necessary insurance authorization, and providing adequate postprocedural observation. Moreover, minor complications that occur in outpatients require the radiologist to act as a primary physician, contrary to the situation for inpatients (10).
The database has permitted us to assess other types of potential stresses on our interventional radiology service. One example is the increasing number of interactions with patients who require special precautions against contagion (11). Treatment of these patients increases room turnaround times and slows morning inpatient rounds.
Another example is the demand for portable procedures performed at the bedside. A separate retrospective analysis performed from January 1991 through December 1996 by using lists prepared from the database and other departmental information sources showed that 4.6% of all procedures had been performed outside of the department (Hahn PF, unpublished data, 1997). Because U.S. Health Care Financing Administration rules require participation and physical presence of a staff radiologist for every procedure (12), portable procedures remove a staff member from the department for 15 minutes to several hours. Moreover, setup time for portable procedures is longer than for routine procedures and includes the time needed to obtain proxy consent for critically ill patients and the time to assemble equipment at the bedside. Although no trend analysis is available, the demand for portable procedures has not diminished as a proportion of inpatient activity.
Most of the individual types of procedures have increased in frequency during the 10 years of data collection. The salient exceptions were pancreatic biopsies and both biliary drainages and biliary procedures as a whole. Pancreatic biopsies showed a decreasing trend that was not statistically significant. The decrease in biliary drainages and ancillary procedures was statistically significant (P < .02). Greater reliance on endoscopic biliary drainage (13) and endosonographic biopsy (14) can account for these observations. The decreasing incidence of biliary procedures reduces the opportunities for radiologists to be trained for interventions of this type.
One hundred forty individuals performed more than 18,000 interventional procedures during the 8 years when the names of operators were recorded for each procedure. In October 1990, five staff radiologists were responsible for the gastrointestinal special procedures, with two more contributing to the GU procedures. In September 2000, there were 11 staff radiologists assigned to perform interventional procedures, all of whom except four had only part-time clinical responsibilities or spent time in other divisions of the department.
Clinical responsibilities of all the abdominal radiologists include substantial diagnostic work in addition to interventional caseloads. Four senior radiologists performed or supervised between 2,000 and 3,000 cases in the past 8 years covered by the database, accounting for slightly more than half (55%) of the procedures. Thus, nearly half of all the procedures were performed by radiologists with a relatively shorter time commitment to special procedures. The large number of cases has spread the procedures to a widened circle of radiologists, including many junior staff. The larger caseload permits training of more radiologists. However, the decrease in biliary interventions requires that senior interventional personnel remain available to consult with their junior colleagues for challenging cases.
The benefits of maintaining the database, such as facilitating research and interstaff communication, quality control, and improved billing reliability, were anticipated at its inception. The database can also be used to monitor equipment, scheduling, and staffing needs, an unanticipated benefit. For example, we used the database to recognize the increased incidence of cases performed on Friday, a phenomenon recently recognized by others (15). Our realization that some cases seemed more urgent at the end of the week permitted us to modify scheduling procedures. Analysis of trends in the type of procedures performed indicates an increasing need for US capacity for liver and thyroid biopsies, while traditional fluoroscopic procedures such as interventional procedures involving the biliary tree are stable or even decreasing.
In summary, we have analyzed the contents of a 10-year database of nonvascular interventional procedures. The clinical community has become increasingly reliant on nonvascular special procedures performed by radiologists. Clinical acceptance has come at the potential cost of increasing stress for the radiologists who perform nonvascular interventional procedures.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, P.F.H.; study concepts, P.F.H., P.R.M.; study design, P.F.H.; literature research, P.F.H.; clinical studies, all authors; data acquisition and analysis/interpretation, all authors; statistical analysis, P.F.H.; manuscript preparation and definition of intellectual content, P.F.H.; manuscript editing, revision/review, and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. M. Bell and D. A. Redelmeier Enhanced weekend service: an affordable means to increased hospital procedure volume Can. Med. Assoc. J., February 15, 2005; 172(4): 503 - 504. [Full Text] [PDF]
|
|
|
|
 |
|
 W. W. Mayo-Smith, M. V. Jayaraman, R. S. Han, D. E. Dupuy, and J. S. Movson Multiinstitutional Computer Database for Recording Nonvascular Imaging-Guided Interventions Am. J. Roentgenol., December 1, 2003; 181(6): 1491 - 1493. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
