Resident On-Call System

Institutional review board approval for this study was obtained before data collection. At our institution, a single on-call junior resident (PGY 2, 3, or 4) is responsible for all consultations, admissions, management of active current inpatients (admitted to the floor or intensive care unit), and any necessary surgical procedures (bedside or those performed in the OR) that take place at our 4 downtown teaching hospitals: a university hospital (UH), an adult level 1 trauma center (ATC), a children’s hospital (CH), and a Veterans Affairs Medical Center (VAMC). If surgical procedures are required, consultations and other patient management duties are done afterward (pages and phone calls are still returned during surgery) unless immediate attention is needed, at which point the backup on-call resident is notified and assists until the junior resident is ready to resume his or her duties. The backup and double-backup on-call residents also cover concurrent emergency operations or consultations, if necessary. However, operations covered by the backup residents, which were few in number, were not included in this analysis because the primary author was not involved.

The amount of time dedicated to on-call activities varies by training level, with so-called primary call performed by PGY 2–4 residents and secondary and tertiary backup call by PGY 4–7 residents. Primary call is front-loaded with the highest number of assigned days during PGY 2; progressively fewer calls are taken in PGY 3 and PGY 4, with no further primary call duties from PGY 5 onward. The term “on call” refers to either a day-call 10-hour shift (07:00 to 17:00), an overnight weekday 14-hour shift (17:00 to 07:00), or a weekend or holiday 24-hour shift (07:00 to 07:00 the next morning).

Database Creation and Economic Analysis

The primary author prospectively kept a call log each time he was on call from July 1, 2014, to June 30, 2016. Among other variables described in our previous work, potentially billable services—such as admissions, consultations, operative interventions, and the most common neurosurgical bedside procedures (external ventricular drains [EVDs] and intracranial pressure [ICP] monitors)—were recorded. ¹³

Patient evaluations and indirectly and directly supervised procedures performed by residents while on call were matched to work relative value units (wRVUs) by using the Current Procedural Terminology (CPT) coding system. The following CPT codes were used: 99223 for history and physical, 99254 for inpatient consultation, and 61107 for both EVD and ICP monitor placement. Bedside procedures were typically done independently by the on-call resident.

For directly supervised operations, only the primary CPT code was used, but no additional codes for each procedure. Reimbursement rates for wRVUs are widely variable; therefore, the published 2014 Medical Group Management Association (MGMA) nationwide median for neurosurgeons of $84.37 per wRVU was used in our data analysis. ¹⁴ Emergency operations were performed under direct supervision; that is, major operations were performed with the on-call attending surgeon physically present and with the junior resident functioning at least as a first assistant if not doing the majority of the case. Given that the attending surgeon must be present for the critical portions of the operation, we elected to qualify the activity of the on-call resident neurosurgeon during directly supervised procedures by using an 80 modifier. By using this modifier, an assistant surgeon is able to bill for 16% of the total cost of the operation. This yields a conservative estimate of the actual work value of the resident during such procedures.

Consultations and Admissions

The primary author’s call volume in terms of days and hours over 2 years is summarized in Table 1. During this time, he evaluated 1929 new patients in consultation or as an admission to the neurosurgical service while on call. Of these, 284 were admitted to the neurosurgery service and 1645 were seen as consultations. The breakdown of where the admissions or consultations occurred is as follows: UH (n = 886, 46%); ATC (n = 717, 38%); CH (n = 267, 14%); and VAMC (n = 59, 3%). ⁸

Using the CPT codes 99223 for history and physical and 99254 for consultation resulted in a total of 6508 wRVUs produced by a single on-call resident for these services during the study period, or 3254 wRVUs per year.

Extrapolating these data to estimate the total number of consultations and admissions seen by on-call residents at our institution during a 1-year period yields the following: 4341 from July 2014 through June 2015, and 4989 from July 2015 through June 2016, or 9330 for the 2 years of the study. ¹³ This equates to 31,442 wRVUs from all on-call consultations/admissions by on-call neurosurgery residents for those 2 years, or 15,721 wRVUs per year.

Indirectly Supervised Procedures

Procedures performed on an emergency basis while on call or within the first 24 hours were also recorded. Neurosurgical intervention performed within the first 24 hours of consultation occurred in 330 (17.1%) patients: 221 (11.4%) interventions from major operations, 69 (3.6%) from EVDs, and 40 (2.1%) from ICP monitors. Indirectly supervised procedures were defined as EVD and ICP monitor placement because these generally occur with the attending physician immediately available, but not necessarily at the bedside. These procedures yielded a total of 544 wRVUs produced during PGY 2 and PGY 3 by a single on-call neurosurgery resident, or 272 wRVUs annually.

Extrapolating these data for the entire cohort of primary call takers in our program results in 527 EVD and ICP monitor placements, which generated 2630 wRVUs using the CPT codes previously mentioned. Thus, a conservative total of the economic output of the indirectly supervised on-call activity of the residents in our program (consultations/admissions plus EVD/ICP monitors) for the 2 years of the study was 34,072 wRVUs; annually this is a production of 17,036 wRVUs.

Directly Supervised Procedures

Directly supervised procedures were defined as emergency operations performed during on-call hours with the attending physician physically present. There were 221 operations and procedures included in this category. Using only the primary CPT code for each operation as well as the CPT modifier 80 resulted in 1120 wRVUs for a single neurosurgical resident during the 2-year study period, or 560 wRVUs annually.

Extrapolating these data to encompass the activity of the total resident call pool results in 1082 directly supervised operations during the study period, amounting to 5478 wRVUs over 2 years, or 2739 wRVUs annually.

Total On-Call Neurosurgery Resident Economic Output

Combining the total number of wRVUs from admissions, consultations, and procedures performed under direct or indirect supervision gives the total theoretical economic productivity of a neurosurgical resident in our program during on-call shifts over the 2-year study period (Table 2). We calculated that a single resident generated a total of 8172 wRVUs over 2 years of on-call shifts, or 4086 wRVUs annually. Extrapolating our data to represent the estimated total financial impact of our on-call residents yields a total of 39,550 wRVUs over 2 years, or 19,775 wRVUs annually (Table 3). These data represent only activities performed while on call and do not reflect any potentially billable activities performed during regular working nonpager hours, thus underestimating the true financial value of neurosurgical residents.

Potential Billing From On-Call Resident Activities

Despite wide regional variation in reimbursement per wRVU, we sought to calculate a conservative estimate of potential billing from resident activities performed solely while on call. We used the 2014 nationwide MGMA median neurosurgery reimbursement rate as a proxy. Over the 2-year study period, indirectly supervised bedside procedures (EVDs and ICP monitors) yielded 544 wRVUs, or 272 wRVUs annually, resulting in $45,890 of potential billing. There were 284 admissions and 1645 consultations, yielding 6508 wRVUs, or 3254 wRVUs annually, resulting in a potential billing of $274,552 per year. There were 221 operative cases performed under direct supervision with an attending physician, yielding 7002 wRVUs, or 3501 wRVUs annually. Using the assistant surgeon billing schedule of 16%, this equals $94,520 of potential billing annually. The total of all these potentially billable services is $344,757 annually for a single on-call neurosurgery resident. Thus, we submit that the annual cost to train a neurosurgery resident appears to be far exceeded solely by on-call productivity. ⁵ Extrapolating our data to reflect potential billing from our entire cohort of on-call residents yielded 39,550 wRVUs over the 2-year study period, or 19,775 wRVUs annually, which equates to $3,336,772 in potential billing over the 2-year period, or $1,668,386 annually.

Quantifiable on-call value is challenging and varies across specialties. A value study of a plastic surgery resident consultation service at a large academic medical center with an ATC evaluated 2367 patients in consultation and found that 10,287 wRVUs could have been generated from the aggregate resident clinical services over the 1-year study period. In their study, more than half of the encounters (52%) resulted in at least 1 procedure, accounting for 3316 wRVUs. This is in contrast to our study, in which nearly twice the number of on-call patient encounters (4665 vs 2367) resulted in one-third of the number of emergency procedures or operations (17% vs 52%). ⁸ A similar prospective multicenter study of 4 resident orthopedic on-call services at 4 tertiary academic medical centers found that 9142 wRVUs could be generated collectively over a 90-day period by these on-call orthopedic services from consultations and performing procedures. ¹⁰ Our study suggests that the value of services provided by on-call neurosurgery residents is quite significant and far exceeds the cost of training with respect to GME funding.

Limitations of the Study

This study resulted from a single neurosurgical resident quantifying his on-call experience, and as such the data are limited to that lone resident. However, it is reasonable to assume that other junior residents at our program would have had similar experiences. Our results are a reflection of the resident call schedule at our institution and may not apply to other programs for which the size of the resident cohort and on-call workload may be different. These results describe the economic value of a neurosurgical resident physician and are not generalizable to other medical or surgical specialties.

Another limitation of our study is that it takes into account only services performed while on call and does not include day-to-day activities, such as rounding on patients, outpatient clinics, and assisting in scheduled operations during regular business hours. Furthermore, our use of the 80 modifier as a means to generate wRVU and dollar value for directly supervised surgeries performed by the on-call resident is not accurate and, in fact, represents a conservative assignment. In many cases, with the exception of pediatrics, the on-call resident does much of the operation with limited direct attending involvement. Thus, our data significantly underestimate the true overall work and economic value of a neurosurgical resident physician.