Joint Hearing of the Committee on Homeland Security and Governmental Affairs of the U.S. Senate and the Committee on Veterans’ Affairs of the U.S. House of Representatives at 1:00 p.m. CDT.
Witness Testimony of Larry H. Conway B.S., R.R.T., Director of Communications, National Association of Veterans Affairs Physicians and Dentists
Mr. Chairman and distinguished members of the Subcommittee:
I am Larry H. Conway and I am the Director of Communications for the National Association of Veterans’ Affairs Physicians and Dentists (NAVAPD) and I am honored to have this opportunity to represent NAVAPD in that role before the Subcommittee. I also currently serve as the Chief of the Respiratory Therapy Subsection at the Washington DC VA Medical Center, and for 38 years have practiced as a respiratory therapist in various hospitals, primarily in management roles. In these roles, I have become extremely familiar with using and developing various methodologies of assessing healthcare staffing needs and productivity systems. NAVAPD President Dr. Samuel Spagnolo regrets being unable to participate today but has asked me to present NAVAPD’s concerns and thoughts on developing a methodology for determining VA’s physician staffing needs, and the VA’s ability to adequately meet patient needs in an efficient, effective manner.
NAVAPD’s focus since its inception in 1975 has been promoting and supporting the highest quality care for our Nation’s Veterans, and caring for those who provide care for them. To that end, NAVAPD supports the development of a balanced, fair and appropriately administered staffing and productivity system that will help assure appropriate staffing levels to provide the excellent care due our Veterans without undue or inequitable stress upon the caregivers. The absence of such a VA-wide system, and the flaws in the systems currently in use in some facilities, have led to productivity assessment approaches that are neither fair nor balanced, and in fact misleading and useless in determining staffing needs and performance levels.
We became aware of concerns about these issues over the last two years through comments from our members. We reviewed the OIG Audit of Physician Staffing Levels for Specialty Care Services (December 27, 2012) and found that it confirmed many of the issues that had been brought to us. The processes being used, where and when used, are fundamentally flawed, based upon the wrong measurement units, and in some cases favored certain staff members while harming or diminishing others. The system can make a physician who performs procedures continually for their entire shift appear less “productive” than a fellow physician who performs procedures only a few hours out the shift. Whether this is because of a lack of understanding of the fundamentals of a staffing and productivity system or intentional, cannot be firmly ascertained. Regardless, these concerns and review of the OIG Audit culminated in an article in the current NAVAPD Newsletter. This article was written and planned for publication before NAVAPD became aware of this hearing and details many of the experiences of NAVAPD members and the parallel findings by the OIG.
Developing such a system for the VA is a challenge, but it is not nearly impossible and should not take a decade to accomplish. In my management roles across the United States, I have devised, reviewed, developed and refined multiple staffing and productivity systems. Developing a system is not complex, though it can be tedious. One barrier to progress is the assumption of an excessive degree of complexity. Beyond selecting the correct measurement units, the greatest difficulty will be in gaining consensus on the application of those measurement units and the assignment of measurement units to various procedures. NAVAPD assigns no blame to the VA for these difficulties and does not seek to engage in controversy or confrontation with the VA. Rather, NAVAPD would like to offer its thoughts to the Subcommittee regarding the misjudgments in developing a system, and further to offer assistance and expertise on how to actualize a viable, beneficial and transparent system well within the time frames recommended in the OIG Audit.
The Fundamental Problems
There are three issues at the heart of the current gridlock of defining and operating a valid system for the VA:
1. Misunderstanding or misconstruction of the basic unit of measurement, the Relative Value Unit (RVU); inclusion of extraneous factors in the RVU.
2. Adding skill-set, procedure difficulty, and stress factors to the RVU. This is a matter of skill-mix, which differs from basic staffing levels
3. Confusing and mixing staffing needs assessment, productivity assessment, and benchmarking.
RVU Selection, Definition, and Construction:
The OIG Audit stated:
“An RVU is a value assigned to a service (such as a medical procedure) that establishes work relative to the value assigned to another service. For example, a service with an RVU of “2” accounts for twice as much physician work as a service with an RVU of “1.” It is determined by assigning weight to factors such as the:
• Time required to perform the service
• Technical skill and physical effort
• Mental effort and judgment
• Psychological stress associated with the service and risk to patient”
With respect, this is precisely the wrong approach and is at the heart of the confusion and disarray of the current system. When asking how many staff members are needed to effectively and safely perform a projected workload, it is an issue of time, not difficulty or skill or physical effort or difficulty or stress. For one thing, a more difficult, more stressful, more skilled procedure will by its nature take longer than a simple procedure
For purposes of determining the total number of staff hours (staffing) needed to accomplish a given workload, the RVU should be a simple, one-dimensional (single-factor) time-based unit. The RVU can be defined as any convenient standard block of time, i.e., one (1) minute, fifteen (15) minutes, one (1) hour, or any block of time that conveniently fits the overall duration of procedures. The VA could and should certainly set a system-wide RVU of perhaps thirty (30) minutes. This will make the data from different services, facilities, and VISNs easy to assimilate, aggregate, and compare without the need for translation of base units.
Unfortunately, the RVUs being used unevenly throughout the VA include all of those factors described in the OIG Audit. They are similar to the Medicare-derived Resource Based Relative Value Units (RBRV). While similarly named, RBRVs and RVUs are not the same and not interchangeable. The RBRV is used to determine the dollar value (reimbursement) of various procedures, and thus includes all of the non-time factors identified above.
For a measurement unit intended to determine the dollar value for a given procedure, as the Medicare RBRV is, inclusion of all of these factors is valid. The impact of inclusion of these non-time factors is illustrated in Figure 1, above.
However, a measurement unit intended to determine just the number of needed staffing hours (which translates to FTEs) should consider only the time for appropriate and safe completion of the projected workload. One-half hour of neurosurgery and one-half hour of wound suturing in the Emergency Department do not require the same skill-level and are not equally difficult. They are thus assigned differing dollar values. But they both take one-half hour of staff time, which is the question when determining how many staff members are needed to complete a whole mix of various procedures.
The question of how many of each type of staff is needed (skill-mix) can be addressed in one of two ways, as described in the next section, but must not be mixed into the RVU.
Assessment of Skill-Mix Need:
As used in these comments, skill-mix means how many staff of various levels of skill is needed. Obviously, a hospital cannot function with only one skill-level or specialty of physician. Having calculated how many total minutes or hours (which can all translate to FTEs) of personnel are needed for all procedures, how does one determine how many Family Practice, Emergency Care, Neurosurgeons, Cardiologists, etc. are needed within that total staffing complement?
The simplest way is to continue to use the RVU as defined previously, but segregate the types of procedures by specialty or skill-level. Thus, the procedures (and associated RVUs) done by Neurosurgeons will be totaled for Neurosurgeons. Those for Cardiologists will be totaled for Cardiologists, and so forth. This process will produce subsets of RVUs for each specialty/skill-level, which will define how many of each specialty/skill-level is required for the projected workload. All of the subsets added together will provide the total staffing complement. This concept is illustrated below in Figure 2.
Alternatively, the skill-mix need can also be calculated by using the RBRV or another unit that considers the factors listed in the OIG report. However, this requires an additional set of calculations and a conversion process between RBRVs and FTEs. There is no significant benefit in this additional, parallel system. Therefore, for the purpose of determining total FTE need and skill-mix need, a one-dimensional time-based RVU is the appropriate tool, not a multidimensional construct like the RBRV.
Confusing Staffing Needs Assessment, Productivity Assessment, and Benchmarking:
Assessing staffing needs and assessing staff productivity are related but not the same, and confusing the two into one system will degrade the effectiveness of the system for both. It will also create a disincentive for staff to participate in either system.
A (relatively) simple means of determining total staffing need and skill-mix has been described.
A productivity system functions in the opposite fashion from a staffing needs system. It should compare the number of staff hours available to the amount of work accomplished. Thus, if there were 1,000 hours of staff time available (based upon a needs assessment) but only 823 hours of work were accomplished (as calculated by RVU), the staff would be considered to be 82.3% productive. The level of productivity can be impacted and made difficult to accurately assess by several factors, some of which are described below.
Factors which can vary facility to facility:
1. Number floors to be covered
2. Acuity of the patients
3. Number, speed, and reliability of elevators
4. Age and speed of equipment
5. Computer systems
6. Number and efficiency of support staff
7. Number of available exam rooms
8. Delays in obtaining a bed
9. Patients not available
10. Teaching obligations
11. Untracked responsibilities such as telephone consults, hallway consults, prep time, documentation.
Fatigue and Delay factors must not be forgotten in determining productivity, while they are often ignored in calculating staffing levels. No one can function at 100% productivity continuously, either for individual health or fatigue reasons, or for the reasons listed above. Productivity specialists consider 5% to 7% a reasonable estimate/allowance for Fatigue and Delay.
Non-tracked responsibilities or obligations that are not directly related to procedures diminish productivity if not considered within the build of the productivity system. Because hospitals tend to build documentation systems around “billable items,” or easily identified procedures, non-billable items are often not counted and thus unavailable for consideration unless recorded manually.
The impact on productivity of resident training is a particularly large factor that is missed in staffing and productivity systems. Even the OIG Audit underplays the impact of teaching. It states:
“VHA officials were also concerned that its National Patient Care Database did not capture all of the physician workload....For example, VHA officials told us that physicians who supervise residents accomplish less workload than their peers who do not supervise residents because the residents will get credit for the work completed. While this may be valid if VHA is trying to establish individual physician productivity, it is not a valid concern when developing a productivity standard for a specific specialty within similar medical facilities.”
In fact, these teaching obligations and the impact upon the entire specialty and facility are significant. Many VHA facilities have specific contracted obligations to use and train residents. Resident training is time-consuming and can reduce significantly an attending physician’s case output or require the physician to spend more hours discharging the same caseload. The more conscientious the teaching, the greater is the impact. Such obligations must be considered when setting staffing levels, productivity factors and goals whether facility or individual focused.
Poorly defined “Encounters” measure used by the VA are defined more in terms of complexity than time, making it difficult to use “encounters” as a denominator to establish staffing need or productivity. The amount of time required varies widely from one encounter to the next, but all are counted as “1.” The more nebulous the measurement unit or documentation unit in terms of time required, the more difficult it is to truly assess staffing needs or productivity of existing staff.
However thoroughly and well consider, a productivity system inappropriately built upon a multifactor measurement unit, like the RBRV discussed previously, can cause hard working and diligent physicians to appear less productive than fellow physicians who do fewer, heavier weighted procedures. This effect is demonstrated in Figure 3 below, which is based upon the assumptions in Figure 1:
Given that the types of procedures done by physicians are not necessarily their choice, but assigned, the situation can arise in which a physician, by virtue of their assigned procedures, could never achieve high productivity in a system that weights by skill and difficulty factors as well as time. A radiologist who is constantly assigned to read chest x-rays could read far more films and work far more hours and never generate the total number of RBRVs as a radiologist who does Brain MRIs or Radio-ablations. If RBRVs are then assumed to equal productivity, the radiologist who is assigned largely chest x-rays will always appear less productive, even if that is not the case.
If productivity as determined by RBRVs is a major determinant in performance assessment and performance pay, the radiologist who is assigned mostly chest x-rays is at a continuous – and perhaps intentional – disadvantage
In a medical system focused on profit, assessing the value of a physician based upon the “production” billable revenue of one versus another might make business sense. In the VA system, profitability is not a factor and so assessing the productivity of a physician should be based upon the time spent producing care results.
Finally, a benchmarking system compares performance on a “select group” of procedures or services that are thought to be highly representative of work associated with and in common with each of the various participating facilities. Of major importance is noting that a benchmark system makes no attempt to account for all procedures or work performed. It therefore does not provide any estimate of the TOTAL work performed in any facility. It is a comparator system and presumes that if a facility has the best profile on the reported procedures, then that facility performed better overall than the other participating facilities.
Benchmark systems are often misused by trying to treat them as productivity systems. The two are completely different and distinct. There is no way to accurately assess true productivity (work produced per staffing unit) unless all work and all staffing is accounted for. By definition and practice, a benchmark system does not account for all of either.
On page 4, the OIG Audit discusses an attempted benchmark looking at infectious disease care and endocrinology care. The OIG investigators then ran productivity comparisons of the two specialties in two different “1a” facilities. While the results imply that one facility was far more productive, the fact is that other procedures and factors not in the scope of the benchmark reporting likely account for some of the variability. This attempt at using benchmark data to derive productivity information produced data that truly only showed that the results were suspect because no standards of measurement and comparison had been established. Because all factors and procedures are not included in a benchmark system, there is little chance of deriving generalized productivity information from it.
A benchmark system may be an effective tool for identifying best practices only if the scope and limitations of its data pool are recognized and considered in any conclusions.
The Greatest Barriers
The greatest barrier to the development and implementation of an accurate Staffing Needs Assessment system and a Productivity Assessment tool will be defining the measurement unit and applying it to all procedures. This will require two major accomplishments:
1. A complete inventory of procedures, events, obligations that account for sizable portions of staff time, billable or not, linked to a procedure or not; and
2. Consensus on the application of the measurement unit to each item in this inventory. For example, getting agreement on “What is the most accurate average time required to perform a Brain MRI?” What is the most accurate average time required to read an EKG?”
The next greatest barrier will be getting staff participation. No one likes another person monitoring them and their work. Health care providers are especially suspicious of such a system. They realize that they are working with people, not building cars, and that “cookbook” approaches do not account for the variability of people and their medical responses. It will therefore be important that the construction of the system and the operation of the system is transparent to all.
Finally, developing a means of easily collecting the data will be key to success. A process that would auto-populate a procedure tracking and counting system will assure the most accurate reporting.
I would like to conclude by reiterating that NAVAPD recognizes the enormity of establishing a Staffing and Productivity system for the VA, but supports that effort, and offers its assistance in making such a system a reality soon. Mr. Chairman, I would like to thank you and the members of the committee for your kind attention. I would be happy to answer any questions from you or other members of the committee.