Some anesthesiologists avoid provision of obstetric analgesia services (OAS) because of low reimbursement rates for the work involved. This study defines the manpower costs of operating an OAS in a tertiary referral center and examines reimbursement for this cost.
The time spent providing OAS in a total of 55 parturients was studied prospectively using a modification of classic time and motion studies.
Mean duration of OAS in our population was 412 +/- 313 min. Mean bedside anesthesia staff time was 90 +/- 40 min, and mean number of visits to each patient's bedside was 6.3 +/- 2.0 visits. Assuming staffing on demand for service (intermittent staffing), a minimum of 2.5 full-time equivalent (FTE) attending anesthesiologists was required to meet demand. With intermittent staffing, labor cost was $325 per patient. Actual practice at Duke University Medical Center is around-the-clock (dedicated) staffing, which requires 4.4 FTEs at a cost of $728 per patient. Neither average indemnity reimbursement ($299) nor Medicaid reimbursement ($204) covered the cost per OAS patient. Breaking even is possible under indemnity reimbursement because operating room reimbursement subsidizes OAS costs. Breaking even cannot occur with Medicaid reimbursement under any circumstances.
Obstetric analgesia services requires a minimum of 2.5 FTE attending anesthesiologists at Duke University Medical Center. With the current payer mix, positive-margin operating room activities associated with the obstetric service are not sufficient to compensate for the losses incurred by an OAS. Around-the-clock dedicated obstetric staffing (4.4 FTEs) cannot operate profitably under any reasonable circumstances at our institution.
PROVISION of an obstetric analgesia service (OAS) is sometimes perceived as a costly, labor-intensive, and poorly rewarded task. 1,2The greatest direct expense in providing the service is anesthesia staff. Obstetric anesthesia workload is unpredictable, which makes forecasting staff needs difficult. We undertook this pilot study to identify the manpower costs to the anesthesia providers of operating OAS. A method to define the labor costs of operating OAS would aid in predicting required staffing versus reimbursement to anesthesia providers. This would be useful both for bidding on OAS and negotiating with hospitals to provide this service. Furthermore, analysis of the third-party payer mix, volume of service, and relative Medicaid reimbursement would demonstrate whether this service can break even in a given hospital setting. At Duke University Medical Center (DUMC), both resident anesthesiologists and certified registered nurse anesthetists (CRNAs) are employed by the hospital, not the attending anesthesiologist’s group practice. The hospital’s revenues from Medicaid cannot be specifically related to its technical and hospital personnel costs associated with the OAS. This study examines the costs associated with providing attending anesthesiologist coverage for on-demand (intermittent) obstetric services and also for around-the-clock (dedicated) service to obstetrics. Both OAS and operating room (OR) activities in the obstetric suite were included in the analysis to determine whether positive-margin OR cases could subsidize OAS.
OAS Time Study
Institutional Review Board approval was obtained at DUMC. The time spent managing labor epidural catheters in a total of 55 parturients was studied prospectively. We used a modification of classic time and motion studies developed by Frank Gilbreth for industry 3and later adapted to hospital work by Lillian Gilbreth and Ruth Kuehn. 4This technique develops labor standards using a series of timed observations. The modified version of this technique requires that the staff performing the task record their own times for activities. This modification has been widely used to develop standards in the hospital industry for nursing care and can be applied to physician services. 5Feldman et al. 6showed that physician estimates of time spent performing patient care responsibilities were reliable. Franzini and Berry 7used self-reported times to determine costs of an anesthesiology residency program. All time spent at the patient’s bedside was recorded by the individual anesthesia staff, including: (1) preoperative evaluation of the chart; (2) directed physical examination and discussion for informed consent with the patient and any family members present; (3) positioning for the block, ascertainment of satisfactory analgesia, and documentation of the procedure; (4) reevaluation of analgesia, top-up doses, or refilling of syringes during the block; (5) discontinuing the catheter and postprocedure instructions after delivery; and (6) postprocedure visit the following day to assess the site and evaluate for possible anesthesia complications.
Times recorded reflect all nonoverlap times spent at the patient’s bedside by an anesthesia care provider in the Labor and Delivery Unit. Nonoverlap times meant that if two providers were present at the same time, only one provider’s time was counted. A second- or third-year resident, an attending anesthesiologist, and a CRNA were dedicated to obstetrics each weekday and weeknight. Weekends were staffed with a resident and an attending anesthesiologist. Staff recorded all time spent in attendance at the bedside of all patients who requested labor analgesia during the fiscal period from August 15–31, 1995. During this period, a total of 74 epidural catheters were placed for 114 deliveries in our facility. Data on 17 patients eligible for inclusion in the study were either incomplete or not collected (23%), primarily during high workload periods when paperwork was deferred for hands-on patient care. Review of age, weight, height, parity, and duration of labor in these patients found no significant differences between excluded patients and the study group.
In our institution, there are three ORs in the Labor and Delivery Unit for obstetric procedures, including cesarean section (CS), postpartum bilateral tubal ligations (BTLs), cervical cerclage placements, and dilation and curretage (D&C) procedures. Time measurement for management of OAS stopped at the point that the decision to perform CS was made in patients delivered surgically. Time in attendance for OAS included time spent at the bedside of patients delivered with vacuum assistance or outlet forceps (operative vaginal delivery). Once a decision to perform CS was made, time spent with the patient was counted as OR time instead of OAS time. OR time included patient preoperative preparation, transport by anesthesia staff, surgical time, attendance in the recovery area, and subsequent postoperative visit.
For clarity, this article examines only the attending physician costs of OAS and obstetric OR activities, without including dependent providers, supplies, or equipment. To determine labor costs, the mix of patient care provided by attending physicians, residents, and CRNAs was determined. During the study period, 42 of 55 labor epidural catheters were placed by the resident on duty supervised by the on-site attending anesthesiologist (76%), 7 were placed by the CRNA supervised by the on-site attending anesthesiologist (13%), and 6 were placed by the attending anesthesiologist without assistance (11%). Of these 55 labor epidural catheters, 13 (24%) were subsequently used to provide anesthesia for CS.
In addition to 13 CSs after labor, there were 10 scheduled CSs during the study period, for a total of 23 CSs. Other OR procedures included 10 BTLs and one postpartum D&C, for a total of 34 OR cases. Total OR time was 45 h and 17 min during the study period. Anesthesia for OR procedures was provided by the attending anesthesiologist supervising the resident in 32 cases (94%) and supervising the CRNA in 2 cases (6%).
Data were entered into a computer spreadsheet (Microsoft Excel; Microsoft Corporation, Redmond, WA) for analysis. Statistics were derived using a package developed by SAS Institute (Cary, NC) for microcomputers.
Patient variables are presented in table 1, and bedside time by procedure type is presented in table 2. Of the 55 parturients studied, 25 were nulliparous (46%), and 30 were parous (54%). During the study period, 60% of deliveries occurred to parous women at DUMC. Labor was induced in 22 of the 55 patients (40%). During the study period, 22% of deliveries were induced for maternal or fetal indications. Methods of delivery included 34 spontaneous vaginal deliveries (63%), 5 vacuum-assisted deliveries (9%), 4 low or outlet forceps deliveries (7%), and 11 CSs (20%).
Measured Time and Motion Data
The mean ± SD bedside time for performing and managing a labor epidural was 90 ± 40 min. Mean time for positioning for the block, ascertainment of satisfactory analgesia, and documentation of the procedure was 43 ± 22 min. There was a mean of 6.3 ± 2.0 visits to patients’ bedside during continuous lumbar epidural analgesia preoperative and postoperative assessment. There was no association between the number of visits and total bedside time spent by the anesthesia staff. There was a significant difference in the time staff spent at the bedside of nulliparous as compared with parous patients (104 ± 9 vs. 79 ± 5 min;P = 0.025). However, there was no significant difference in the number of visits to the bedside, implying that visits were longer for nulliparous patients. There was no significant relationship between maternal weight and bedside time or number of visits. In addition, there was no significant difference in bedside time or number of visits spent on patients with epidural catheters placed at night compared with those placed during daylight hours.
Staffing Required for Direct Patient Care
It is our practice for the attending anesthesiologist to be present for the placement of all regional anesthetics and continuously present during CS at least until delivery of the infant. Average duration of CS during the study period was 93 ± 31 min. Average duration of BTL during the study period was 66 ± 15 min. Total time spent in direct clinical activities was 156.3 h during the study period (111 h OAS time and 45.3 h OR time). This represents an average of 9.2 h/day of direct patient care to meet the level of demand present during the study period. This would equal 64.4 h in one 7-day period. The average DUMC clinical assignment was 56 h per anesthesiologist per week, requiring 1.15 full-time equivalent (FTE) attending anesthesiologists present to meet the level of demand during the study. Unfortunately, this demand occurs randomly throughout the 24-h day, requiring staff to be available around the clock to provide services when needed. This randomness must be accounted for in staffing for unscheduled cases by setting a productivity target allowance. In addition, paid time off such as vacation, holidays, and sick leave must be accounted for to determine appropriate staffing levels.
Productivity Target Allowance
To determine the lowest labor cost per patient possible, one must look at the scenario of optimal productivity. Because only 0.5% of OAS cases were scheduled in 1998, we chose a target productivity level of 70% for this analysis of OAS staffing. Adjusting for a target productivity of 70% requires staffing 1.5 FTE to meet the demand for OAS.
Paid Time Off Allowance
To staff the 1.5 FTE to provide OAS, allowance for paid time off must be made. A factor of 16.7% (0.5 FTE) was used to account for DUMC vacation, sick, and meeting time off, increasing the 1.5 FTE to 2.0 FTE anesthesiologists required to meet demand for OAS. In addition, DUMC allocated 1 nonclinical day within the 5-day work week to be used for research, teaching, lecture preparation, writing, and other academic pursuits. This resulted in a minimum staff requirement of 2.5 FTE attending anesthesiologists to meet the workload demand for OAS when using the intermittent staffing model (2.0 FTE × 1.25).
The actual practice at DUMC is to staff OAS with one FTE attending anesthesiologist, requiring a minimum of 3.0 FTE to fill the position around the clock. A paid-time-off factor of 16.7% (0.5 FTE) was used to account for DUMC vacation, sick, and meeting time off. In addition, DUMC allowed 1 nonclinical day within the 5-day work week to be used for research, teaching, lecture preparation, writing, and other academic pursuits. This resulted in a minimum staff requirement of 4.4 FTE attending anesthesiologists to fill the one position (3.5 FTE × 1.25).
Average compensation for attending anesthesiologists at DUMC, including benefits and malpractice insurance coverage, was $206,405 (table 3). This is direct physician compensation and benefits and does not include other practice costs and university teaching and research costs (“Dean’s tax”). The annual cost of 2.5 FTE attending anesthesiologists to meet OAS demand was $516,012 per year, or $1,413 per day ($516,012/365.25 days per year). Total cost for attending anesthesiologists during the study period was $24,021 ($1,413 per day × 17 days). The cost per OAS patient during the study period was $325 ($24,021/74), assuming OAS is staffed on an intermittent basis. This model assumes that an anesthesiologist will be available to meet all demands for OAS to receive maximal reimbursement. Staffing OAS with one attending anesthesiologist around the clock, as DUMC does, would yield an actual cost per patient of $728.
Reimbursement for OAS
Directly related to staffing OAS is the issue of reimbursement for services rendered. There are two major payers for obstetric health care at DUMC: (1) indemnity or insurance payers, and (2) the Medicaid program. To determine the labor cost for OAS, one must allocate the total labor costs over the volume of procedures performed. Clearly, a change in the volume of work performed with a given level of staffing will affect the cost per patient. If the service is staffed around the clock, as at DUMC, the labor costs are fixed such that underutilization will result in a higher per-patient cost.
For fiscal year 1998, the average DUMC indemnity reimbursement for OAS was $299 (table 4) compared with the cost of $325 per patient, assuming OAS is staffed on an intermittent basis. Total reimbursement for the study period would be $35,615 if all patients were covered by indemnity policies, compared with staff costs of $24,021. The indemnity reimbursement for obstetric OR cases can subsidize the inadequate indemnity reimbursement for OAS when using the intermittent staffing model.
Medicaid within North Carolina currently reimburses $204 for continuous epidural analgesia during labor and delivery, whether vaginal or CS. If all patients during the study period had been covered by Medicaid, the reimbursement for all 74 epidural catheters for delivery placed during the 3-week period would be $204 × 74 = $15,096. There is no further reimbursement for CS after labor epidural analgesia (i.e. , one price for delivery, whether vaginal or CS). Reimbursement for scheduled CS would be $204 for each of 10 cases, or $2,040. Reimbursement for 10 BTLs would be $1,850. During the study period, no BTLs were performed on the day of delivery using a preexisting epidural catheter, which would have led to a reduced reimbursement of $97 per case. Medicaid payment for CS or BTL under general anesthesia and for D&C and cerclage placement is $0.90/min. One D&C was performed, and reimbursement would have been $70 (table 5). Total reimbursement for the study period if all patients were Medicaid recipients would therefore be $19,153, compared with minimum staff costs of $24,021. Breaking even cannot occur with Medicaid reimbursement under any circumstances (table 6).
Applied to annual procedure volumes, reimbursement for fiscal year 1998 would have been $461,543, compared with staffing costs of $516,012. This assumes all patients were covered at Medicaid rates and staffing was perfectly accomplished using an intermittent staffing model (i.e. , no missed cases or lost reimbursement opportunities).
The additional reimbursement from OR procedures during the study period would be $3,960 (10 CSs at $204 plus one D&C at $70 plus 10 BTLs at $185). Subtracting this reimbursement from the labor cost leaves a net labor cost per epidural of $271 ($20,061/74) at the volume of labor patients and OR cases performed during the study period. The OR cases reimbursed at Medicaid rates do not generate a positive margin to cover the overall cost of providing OAS.
Representative Nature of the Data
Methods of delivery during the study period were similar to DUMC annual rates, with a spontaneous delivery rate of 60%, CS rate of 22%, and 18% assisted deliveries (forceps or vacuum). The national rate of CS was 20.6% of deliveries in 1996. 8Operative deliveries (forceps and vacuum-assisted) are not tracked nationally because of lack of specific coding, but a survey of United States university programs in 1994 showed rates between 5.5 and 20.7%. 9The overall labor induction rate of 22% can be compared to a rate of 22.3% reported by Queen Elizabeth Hospital, South Australia, 10an elective induction rate of 12.3% reported in a U.S. community teaching hospital, 11and a rate of 14% of low-risk deliveries in Toronto. 12The induction rate in England ranged from 17 to 20% of all deliveries from 1989 to 1995, reported to the Maternity Hospital Episode Statistics system. 13
Mean time for positioning for the block, ascertainment of satisfactory analgesia, and documentation of the procedure was 43 ± 22 min. This compares with a previous study of epidural versus spinal anesthesia for CS in which the time from entering the OR to incision was 46 ± 11 min. 14Our average clinical assignment was 56 h per anesthesiologist per week. This is similar to the national average derived in a survey by Abt Associates for the American Society of Anesthesiologists in 1994. ††
Average labor costs per attending anesthesiologist at DUMC was $206,405, including compensation, benefits, and malpractice insurance coverage. Compensation alone was $160,778. Franzini and Berry 7found average total costs per attending anesthesiologist of $336,313 at the University of Texas–Houston. Compensation alone was $134,525. The average costs are not directly comparable because Franzini and Berry included the cost per faculty member of all supporting resources and University “Dean’s tax.” Our study examines the labor costs to the anesthesiologist rather than the total costs to the University. Labor costs may be more easily compared because supporting resources are highly variable from practice to practice.
The cost per patient during the study period was $325, assuming OAS is staffed on an intermittent basis. Dedicated staffing yields a cost per patient of $728. Total reimbursement to the attending anesthesiologist’s practice for the study period if all patients were Medicaid recipients would have been $19,153, compared with attending staff costs of $24,021. Medicaid reimbursement for all obstetric procedures during fiscal year 1998 would be $461,543, compared with attending staff costs of $516,012. If all patients were covered by indemnity reimbursement, there would have been an adequate positive margin to defray the salary and benefits for attending anesthesiologist coverage on an intermittent basis. However, covering the service intermittently increases the probability that the attending anesthesiologist will not be available when demand for OAS arises, thus losing the opportunity for maximal reimbursement.
It is unlikely that any anesthesiologist’s OAS practice is comprised of Medicaid patients exclusively. However, making this assumption allows an examination of whether the Medicaid reimbursement for OAS is sufficient to meet costs rather than requiring indemnity payers to subsidize Medicaid patients. The argument that anesthesiologists will benefit from increased volume of services by performing more obstetric anesthesia services for Medicaid patients does not hold because there is a negative contribution margin for each procedure performed.
Productivity and Breaking Even
In emergency service areas such as the OR or the emergency department, departmental productivity levels of 70–80% are considered optimal because excess capacity must be built in to handle unanticipated cases and emergencies. 4There is an exponential decay in the ability to accommodate new service requests once this threshold is exceeded. The wait for delivery of pain relief can become unreasonably long. Even in ambulatory surgery centers in which few unplanned procedures are performed, a departmental productivity level of 85% is considered optimal. 15,16Assuming 90 min of contact time per labor epidural placement, OAS at DUMC cannot break even (table 6). Very few studies have addressed the issue of productivity of anesthesiologists specifically. Posner and Freund 17recently suggested using billed hours of anesthesia care per clinical day as a measure of productivity for attending anesthesiologists. Productivity rates in this study ranged from 10 to 17 h of billed anesthesia care per attending anesthesiologist per clinical day, with concurrency rates of 1.6 to 2.2 cases per attending anesthesiologist. Concurrency rates may be much higher with the longer duration of labor analgesia, but these cases are not billed by hourly rates. An important note is that the American Society of Anesthesiologists Practice Guidelines suggest that an anesthesia care provider must be in-house during the conduct of a labor epidural. No allowance for this availability was factored into the intermittent staffing analysis. DUMC, like many busy practices, chooses to have dedicated attending presence in obstetrics to meet this need for availability. Productivity as defined by Posner and Freund was not measured in this study.
A recent survey by the American Society of Anesthesiologists revealed Medicaid reimbursement rates from $55 to $169.55 in eight other states. 18Although the $204 reimbursement provided by the state of North Carolina is inadequate to cover the cost of an anesthesiologist’s services in providing OAS, it is nearly twice the average from the other eight states surveyed. These results suggest that anesthesia groups are correct in the perception that OAS is not adequately reimbursed and that anesthesiologists are providing the service to Medicaid recipients at a true loss. The argument that obstetric anesthesiologists should provide services for a lower average salary is unrealistic. The opportunity exists for an obstetric anesthesiologist to move practice to the OR, where reimbursements are higher, and current job market opportunities exist. We now know our true variable cost for providing this service, and this information should be used wisely to price services to third-party payers. Optimization models using directed dependent providers may influence the conclusions presented here. Analyses such as we present are the building blocks of economic survival in today’s medical marketplace. At the time of this pilot study, only continuous lumbar epidural analgesia was offered at our institution. A follow-up study to evaluate the economic effect and manpower requirements attending changes in practice, such as the use of the combined spinal–epidural technique and patient-controlled epidural analgesia, is planned.
We conclude that, even with intermittent staffing, there is a negative contribution margin for anesthesiologists to provide OAS at DUMC. With the current payer mix, positive-margin OR activities associated with the obstetric service are not sufficient to compensate for the losses incurred by OAS. Around-the-clock dedicated obstetric staffing (4.4 FTEs) cannot operate profitably under any reasonable circumstances at our institution.