4

VHA Facilities Management (Engineering) Staffing Methodology

The committee learned that the Veterans Health Administration (VHA) does not have a process to develop an overall staffing model for its Facilities Management (Engineering) workforce and that each Veterans Administration Medical Center (VAMC) uses its own method to define how many workers are needed, including decisions about outsourcing portions of the workload. Chapter 3 of this report describes different approaches to staffing models and important qualities for any staffing model to exhibit, such as validity (accuracy), utility, and scalability. Staffing models are not static or one-size-fits-all. Any model must have the capacity to incorporate parameters¹ that are unique to a given organization or department if those parameters reflect substantial differences in workload requirements. Any model must also be able to adjust the parameters as conditions change over time. For example, changes in technology could impact the workload estimates that underlie the model, perhaps reducing workload in some ways, while increasing it in other ways. The model also needs the capability to address variations across sites.

This chapter describes a staffing methodology that is specific to VHA Facilities Management (Engineering). It focuses on estimating staffing needs at the VAMC level, with the capability to aggregate information at the Veterans Integrated Services Network (VISN) or Headquarters (HQ) level.

As described in Appendix C, VAMCs vary significantly in their clinical complexity and they also vary significantly in their infrastructure complexity. Any effective Facilities Management (Engineering) staffing methodology must be able to reflect these infrastructure complexity differences across VAMCs to the extent practicable. The methodology must also reflect the ways in which different aspects of infrastructure complexity differentially impact Facilities Management (Engineering) functions. For example, the number and complexity of funded construction projects impacts the Capital Projects function much more than the Operations and Maintenance (O&M) function. This chapter describes a rationale for defining the baseline staffing level necessary to address critical Facilities Management (Engineering) functions and responsibilities at a VAMC with a low level of infrastructure complexity, followed by application of several infrastructure complexity parameters that require a variance from baseline staffing levels. Each infrastructure complexity parameter impacts Facilities Management (Engineering) workload differentially across VAMCs, such as the amount of square footage devoted to high-intensity uses like operating suites or emergency care, or the age and condition of facilities. The committee identified many infrastructure complexity parameters in its data-gathering workshops and meetings. The staffing methodology described in this chapter can

___________________

¹ Parameter—a variable entering into the mathematical form of any distribution such that the possible values of the variable correspond to different distributions. As stated in https://www.dictionary.com.

incorporate as few or as many infrastructure complexity parameters as VHA deems necessary. The committee heard from several presenters that including many parameters in a staffing model leads to a highly complex model that is very difficult to estimate and likely too unwieldy to use. VHA will need to carefully choose a reasonable set of infrastructure complexity parameters as it applies the staffing methodology to create a VHA Engineering Staffing Model. Last, the chapter closes with an example illustrating how the VHA Engineering Staffing Methodology could work, using notional infrastructure complexity parameters and staffing variances beyond the baseline.

As noted in the interim report and Chapter 3, any resource planning and staffing methodology requires an estimate of the amount and types of work to be completed and the time and effort necessary to complete the designated work (NASEM, 2019d). As Joseph Crance from the U.S. Air Force (retired) noted, “Manpower forecasting models should be built on an accurate measurement of workload content, i.e., accurate man-hour measurement at the process level” (Crance, 2019, slide 15; NASEM, 2019e). Chapter 3 described a variety of methods that can be used to estimate workload and potential data sources. Some data sources may be useful, but only to one function of the Facilities Management (Engineering) workforce. For example, analysis of time taken to complete work orders could be used to determine workload—but may be relevant only for O&M-type work that is handled through a work order process. It also depends on complete, accurate, and reliable work order information.

Chief engineers, who typically have spent time working at lower organizational levels, and other highly experienced Facilities Management (Engineering) staff may be a source of workload estimates, serving as part of a pool of subject-matter experts (SMEs) who can be assembled and tasked with using their experience to derive estimates of time required for each of many different job duties. Time studies or similar techniques can also provide direct time estimates for tasks independent of work order analysis (NASEM, 2019d). Chapter 5 will provide more specific recommendations about the involvement of VHA SMEs in applying the methodology to build the model.

It is crucial that any approach to estimating workload recognizes that workers do not, nor cannot, spend 100 percent of their work hours performing core work activities. For example, workers are authorized to take leave and to take rest breaks during workdays. Also, workers in any job spend some of their time on activities that are important for the organization but are not their core job duties, such as attending training or team meetings and documenting their work in some fashion. At some locations, there may be substantial travel time between buildings or work sites. The VHA Nurse Staffing Model explicitly recognizes this, as the committee heard from the acting chief nursing officer, Beth Taylor, in March 2019. Presenters from manpower planning (Norman, 2019; NASEM, 2019e) and property management organizations (NASEM, 2019e; Schmeidler, 2019) also made this point. Care must be taken to account for nonavailable time, work-related travel time, and time spent performing noncore job duties, without double counting. For example, if SMEs are asked to judge the amount of time required to perform the duties of a maintenance supervisor, they must be provided specific guidance on exactly what they are being asked to estimate—core and noncore time, including travel between sites or not, and so on. Nonavailable time can be accounted for when turning workload estimates into the number of full-time equivalents (FTEs) required—for example, assuming that 15 percent of the hours available for an FTE will be nonavailable.

LINKING VHA FACILITIES MANAGEMENT (ENGINEERING) STAFFING WITH TARGET LEVELS OF PERFORMANCE

Defining Target Performance Levels for VHA Facilities Management (Engineering)

As noted in prior chapters, to establish a staffing model, the input to the model must be based on shared understanding of the performance level(s) to be achieved, along with at least a general understanding of how performance will be measured. For example, in report on Federal Aviation Administration (FAA) systems specialists (NRC, 2013; hereafter the ATSS report), the key performance measure was the fraction of time that the National Airspace System was fully operational, with 100 percent as the desired performance level or goal. A perfect VHA Engineering Staffing Model would relate the Facilities Management (Engineering) staff required directly to outcomes that matter to VHA patients and to those funding the facilities—that is, Congress. It may be challenging, but should be possible for VHA to translate its core mission of “honoring America’s veterans by providing exceptional health care that improves their health and well-being” into desired performance levels that

are applicable to Facilities Management (Engineering), and can, in turn, be broken down into more specific goals and objectives with associated performance measures for more specific functions—for example, for grounds maintenance or work order data entry.

Desired performance levels can be described as striving to achieve some positive level, such as “showplace facility,” or “100% of preventive maintenance performed at recommended intervals,” or “completing 100% of work orders within a specified time frame.” Desired performance levels can also be described as striving to avoid negative (or “never”) events such as zero unplanned power outages or zero operating room unavailability due to problems with positive air flow or temperature control.

The committee heard that VHA Facilities Management (Engineering) has specific policies for the performance of boiler plants, electrical testing, legionella, vehicles, and energy (Broskey and Alverez, 2018). However, except for an expectation to meet Joint Commission standards, there appear to be no specific requirements for heating, ventilation, and air conditioning (HVAC), plumbing (beyond legionella), carpentry, painting, work order system, engineering administration, or project management. Similarly, the committee heard that there are no specific performance expectations related to use or management of existing work order systems, such as how and when data are entered, by whom, and with what degree of accuracy the information is or should be tracked (NASEM, 2019a). Without a clear set of performance-level requirements, it is not possible to definitively determine appropriate staffing levels. For example, should HVAC terminal units be calibrated monthly, quarterly, annually, according to manufacturer’s recommendations, or not at all? VHA currently lacks clear performance level requirements for maintenance, operations, capital projects, and engineering administration.

One presentation recommended that performance levels for Facilities Management (Engineering) be based on the following two factors: the level of availability that medical leaders desire for different areas within a given health-care facility, coupled with the level of quality they expect for the available space (Walker and Dillinger, 2019). Facilities Management (Engineering) experts can then forecast the level of funding or the level of staffing required to achieve these desired levels, and can also link them with consequences associated with failing to achieve recommended funding or staffing levels. The committee heard that at least some availability data are already collected by the VA—for example, the fraction of time that a patient or member of the medical staff is unable to be serviced due to engineering system nonavailability. Currently, these data sets are not combined with existing facilities and staffing data such as the Capital Resource Survey (CAPRES).

Some presenters suggested using levels of service definitions provided by APPA (formerly the Association of Physical Plant Administrators) for higher education facilities in the areas of maintenance, custodial, or grounds functions (Hunter, 2019; Jackson, 2019). See Figure 4.1 for an example of the APPA levels of service narrative definitions for maintenance. VHA could choose one of these levels as the performance toward which it strives.

In summary, while the committee believes that it is necessary to develop a VHA Facilities Management (Engineering) Staffing Model based on a shared understanding of the desired level of performance for different functions, it could not find evidence that such performance levels have been defined across all VAMCs, beyond those that define requirements for specific systems or processes, or perhaps the implied goal of “achieving or surpassing Joint Commission standards.” A recent Government Accountability Office report (GAO, 2018) reached a similar conclusion:

The committee was informed of numerous areas of Facilities Management (Engineering), such as work order management, that do not have policies defined.

RECOMMENDATION 4.1: The Veterans Health Administration (VHA) should establish and define desired performance levels for VHA Facilities Management (Engineering) that flow from the organization’s mission or policy and that can be used to drive more specific performance goals, objectives, and performance measures for Facilities Management (Engineering).

RECOMMENDATION 4.2: The Veterans Health Administration should establish comprehensive minimum performance levels for all areas of Facilities Management (Engineering) beyond the limited policy currently available.

Defining Key Performance Indicators that Reflect Target Levels of VHA Facilities Management (Engineering) Performance

After performance-level requirements have been established, it becomes possible to identify key performance indicators (KPIs) that measure the status of or progress against them. It is necessary to define desired or required performance levels or targets in a descriptive or narrative form first, as described in the preceding section, but then

it is also necessary to define exactly how the desired levels will be measured and which indicators will be used to gauge whether the desired levels and targets are being met.

Chapter 2 listed a variety of KPIs that are commonly used to evaluate facilities management performance, including many types of KPIs used by facilities management organizations to track and evaluate performance in capital planning, operations and maintenance, and engineering administration functions. These include cost-based KPIs such as annual operating cost or life-cycle cost, efficiency-based KPIs such as work order completion rates, and outcomes-based KPIs such as the occurrence of adverse events. The committee heard from several health-care facilities management and property management organizations that apply a common set of KPIs across all owned facilities and a centralized system for storing and managing the KPI data (see Appendix A for a list of presenters.) Furthermore, the committee heard that VHA Facilities Management (Engineering) performance is evaluated regularly, and almost constantly, by entities such as the Office of the Inspector General, Joint Commission, Occupational Safety and Health Administration (OSHA), Environmental Protection Agency (EPA), Nuclear Regulatory Commission, Veterans Administration Central Office, and VISNs (NASEM, 2019d). The potential issue is that there is no clear linkage to a common set of performance levels established for VHA Facilities Management (Engineering), such that there is clarity about which KPIs matter most and which ones directly flow from the VHA mission statement. For example, the FAA uses an Adjusted Operational Availability metric that is the ratio of time that the national airspace system was fully functional to the maximum amount of time that the national airspace system could have been fully functional in a given period (e.g., a fiscal year). This is not the only KPI that matters, but it is clearly one that is of primary importance to the FAA’s mission and is highly relevant to the core function of system specialists. As the level of KPI aggregation increases to the VISN or VHA level, outcome measures take on more salience—for example, with the ultimate funding source of Congress.

The committee heard about KPIs gathered by the Office of Capital Asset Management, Engineering, and Support (OCAMES) for use in benchmarking performance of Facilities Management (Engineering) departments at the VISN level. These include work order completion, critical utility systems maintenance, and measures that reflect performance on addressing gaps identified in the Facility Condition Assessments, and on nonrecurring maintenance (NRM) and minor construction projects (Broskey, 2019; NASEM, 2019f). Many more measures are needed for a comprehensive picture of engineering effectiveness. Obvious examples are the patient outcome measures, but also patients’ reported experiences and satisfaction. The committee heard that such data are already collected—for example, measures such as the fraction of time that a patient or member of medical staff is unable to be serviced because of engineering system nonavailability. However, these are not combined with the existing facilities and staffing data such as CAPRES.

The ATSS report described a hazard model in which specific negative events are evaluated in terms of frequency of occurrence and the severity of the consequences if they do occur (NRC, 2013). Similarly, The Joint Commission provides a Survey Analysis for Evaluating Risk (SAFER) matrix. If an element of performance at a health-care facility is deemed out of compliance, it is cited as a requirement for improvement (RFI) and then placed in the SAFER matrix to index its criticality. Figure 4.2 is a copy of the SAFER matrix.

The committee also heard about a Risk Identification, Triage, Mitigation, and Sustainment (RiTMS) process that is being piloted by the VHA Office of Occupational Safety, Health, and GEMs Program (Dulaney, 2019a,b). The process is used to assign a level of risk to employee safety at any given facility based on a review of KPIs and an onsite review. An initial pilot test has been completed, and RiTMS is currently being implemented at 32 VAMCs (Dulaney, 2019a,b).

Even gross measures such as counts of adverse events can be KPIs. Both The Joint Commission and the Agency for Healthcare Research and Quality (AHRQ) have developed lists of serious, preventable medical errors that can be used as indicators of hospital performance (e.g., Patient Safety Primer Never Events, published by AHRQ). The same concept could be extended to identify “never events” for Facilities Management (Engineering) performance. Figure 4.3 provides a notional list of critical or “never events” that were mentioned in discussions with chief engineers and medical center directors as potential KPIs (NASEM, 2019a,b,c). The presumed underlying performance level goal is that these should never happen.

Analyzing Links Between VHA Facilities Management (Engineering) Staffing and Performance

Ideally, different levels of VHA Facilities Management (Engineering) staffing could be linked to specific KPIs reflecting target levels of performance. For example, VHA could link Engineering staffing levels with the risk of negative events occurring, after those negative events, in turn, have been linked to patient outcomes. This type of linking would provide a concrete basis for chief engineers and medical center directors to gauge the risk of negative event occurrence associated with allocating less than the requested budget (or FTEs) for Facilities Management (Engineering) and estimate the likely reduction in risk of negative event occurrence associated with allocating more than the requested budget (or FTEs). Absent links between staffing levels and KPIs (such as the risk of negative events) conversations between chief engineers and medical center directors are based on instinct or past experience or generalized degree of risk aversion.

As noted in Appendix C, VHA already captures extensive data about facilities and associated costs, and it could potentially capture even more using big data techniques. The challenge seems to lie in identifying the most critical KPIs for Facilities Management (Engineering), which should flow from a shared understanding about desired performance levels, followed by systematically linking the KPIs with high-quality data about Facilities Management (Engineering) staffing levels and costs, to include costs for contracted labor. If VHA can address this challenge, statistical techniques can be used to examine the links between staffing input variables and relevant KPIs, providing robust tools for forecasting the impact of Facilities Management (Engineering) staffing changes on relevant KPIs. These analyses can be conducted at different levels of specificity—for example, at the department or function level within a VAMC, across VAMCs within a VISN, or even at the national level.

The committee understands that it will take time and resources to achieve a shared understanding of target performance levels across VAMCs, identify high-quality KPIs to measure performance, and achieve credible ways to encode staffing information. While this effort is being undertaken, the committee believes that the staffing methodology proposed later in this chapter can be enacted by VHA. The methodology will rely heavily on expert judgments to start, with decision support from analyses for which high-quality data are already available—for example, examining which infrastructure parameters are most strongly related to current Facilities Management (Engineering) staffing levels. As the methodology input and outputs are reviewed and adjusted over time, decision support based on empirical links between staffing levels and performance will likely increasingly play a prominent role.

RECOMMENDATION 4.3. To the extent possible given the availability of high-quality data, the Veterans Health Administration should use analytic techniques to empirically examine the links between input variables related to Facilities Management (Engineering) staffing levels (e.g., labor hours, labor costs, contracted services costs) and key performance indicators reflecting a shared understanding of desired levels of Facilities Management (Engineering) performance.

STAFFING METHODOLOGY

This section describes the committee’s vision for the VHA Engineering Staffing Methodology. It attempts to balance the need to incorporate critical infrastructure complexity parameters that impact Facilities Management (Engineering) workload and that vary across VAMCs, with the need to create a communicable, transparent model that will be easy to explain and understand, such that chief engineers, medical center directors, and other stakeholders can feel confident about the results it produces. The methodology provides a foundation for establishing and justifying VHA Engineering staffing requests, just as VHA Nursing has a staffing model based on the amount of work required to provide the necessary levels of patient care that translates into the number of nurses required at any given VAMC.

Facilities Management (Engineering) Functions Included in the Staffing Methodology

As noted in Chapter 1 and described in Appendix C, the study sponsor indicated that the following VHA Facilities Management (Engineering) functions should be included in the scope of this study.

• Engineering Administration (EA)
• Capital Projects (CPs)
• Operations and Maintenance (O&M)

Levels at Which the Staffing Methodology Should Be Built and Applied

In the committee’s judgment, the most appropriate level at which to apply a VHA Facilities Management (Engineering) Staffing Methodology is at the VAMC level so that it can be used to inform staffing and budgeting decisions that are the purview of the medical center director at each VAMC. To ensure that the resulting

VHA-designed Facilities Management (Engineering) Staffing Model is consistent across VAMCs, the model should be managed centrally, similar to other VHA staffing models such as the Nursing and Occupational Safety staffing models. Later sections of this chapter describe how the parameters themselves lead to staffing estimates that vary across VAMCs, in systematic and transparent ways.

The VHA Facilities Management (Engineering) Staffing Methodology would not only provide a tool that chief engineers can use to estimate staffing requirements, but it also provides an objective basis for a discussion between the chief engineer and the medical center director about the staffing request. When VHA uses the methodology to create a Facilities Management (Engineering) Staffing Model, it should make every aspect of the model formulation transparent, so it will be easy for medical center directors to understand the basis for the Engineering staffing request, similar to the way the Nurse Staffing Model is used to inform discussions with the medical center director about the Nursing staffing request. If the same methodology is used across all VAMCs, medical center directors can be confident that the methodology used to estimate Engineering staffing at their location is the same as the methodology being applied in every VAMC. If an exception from the standard application of the Facilities Management (Engineering) Staffing Methodology is requested by the chief engineer at a given VAMC, it will be easy for the medical center director to understand that an exception is being requested and to engage in a discussion with the chief engineer about the rationale for the exception.

The committee notes that the VHA-designed Facilities Management (Engineering) Staffing Model output may align with current staffing levels, or it may suggest that current staffing levels are too low or too high at some VAMCs. The goal is to generate a clear and logical basis for estimating Facilities Management (Engineering) staffing needs, which provides a basis for conversations about staffing requirements and, ultimately, the medical center director’s decision about how much funding to allocate for Facilities Management (Engineering) at his or her VAMC. If desired, VHA can compare Engineering model outputs across VAMCs and can also aggregate them for comparison at the VISN or HQ levels.

Establishing Baseline Facilities Management (Engineering) Staffing Requirements

Every VAMC requires enough Facility Maintenance (Engineering) staff to perform duties and meet responsibilities to the levels specified in Joint Commission standards and in VHA policies and directives (Broskey and Alvarez, 2018). The model must also reflect Facilities Management (Engineering) workload requirements associated with the federal government- or VHA-specific purchasing and contracting requirements. The committee notes that many of the aforementioned standards, policies, and directives do not specify exactly how responsibilities and duties should be performed by Facilities Maintenance (Engineering) staff, or by which labor categories or by contracted versus in-house labor, or exactly how many FTEs are required to carry out the work in a given VAMC. The committee believes that it is possible for VHA to define baseline staffing levels and job grade structure required for a VAMC with a low level of infrastructure complexity. Every VAMC must have at least this level of staffing. Beyond the baseline staffing levels, staffing variances reflect additional workload associated with infrastructure complexity parameters that vary across VAMCs. VAMCs with a moderate level of infrastructure complexity require more Facilities Management (Engineering) staff than baseline levels, but not as many as VAMCs with a high level of infrastructure complexity. Staffing variances at facilities with a higher degree of infrastructure complexity could entail adding more positions to baseline levels, but could also involve changing the job grade structure for some positions, or a choice to outsource some of the workload to contractors.

Figure 4.4 provides a potential starting point for a baseline model, illustrating the type of positions likely needed at even a small VAMC with low infrastructure complexity. This figure is adapted from an organizational chart provided by the sponsor, with the numbers of staff indicated for each position removed, but with reporting relationships and position titles retained (Broskey and Alverez, 2018). Note that Figure 4.4 excludes functions that the sponsor designated as out of scope, such as Biomedical and Occupational Safety. The committee believes that the chart provides a reasonable list of positions needed at every VAMC, although it does not specify all of the work that must be performed by VHA employees. The workload associated with some positions could be outsourced if it is deemed necessary or cost-effective to do so (e.g., due to challenges in filling vacant positions). With SME input, VHA Facilities Management (Engineering) can articulate the minimum workload requirements to justify the numbers of staff in the baseline model.

FINDING 4.1: Several Facilities Management (Engineering) positions and roles are mandated by VHA, The Joint Commission, or other federal government requirements, although some of them do not require an FTE at each VAMC.

RECOMMENDATION 4.4: The Veterans Health Administration (VHA) should substantiate baseline, VHA Facilities Management (Engineering) staffing levels, including all mandated roles, common to all Veterans Administration Medical Centers by function (Engineering Administration, Capital Projects, and Operations and Maintenance) and job (e.g., project engineer, maintenance foreperson, skilled trades).

RECOMMENDATION 4.5: The Veterans Health Administration (VHA) should establish baseline VHA Facilities Management (Engineering) staffing levels common to all Veterans Administration Medical Centers by function (Engineering Administration, Capital Projects, and Operations and Maintenance) and job (e.g., project engineer, maintenance foreperson, skilled trades). This baseline model should take into account all mandatory positions and roles dictated by VHA policy and other federal directives, or Joint Commission standards.

Incorporating Infrastructure Complexity Parameters into a Baseline Staffing Model

The baseline staffing levels should be supplemented with a set of parameters that reflect significant ways in which infrastructure complexity varies across VAMCs and thus impacts the workload for Facilities Management (Engineering) departments. The parameters can be reviewed on a regular basis—for example, annually—and adjusted as needed to account for changing conditions. Each parameter should be accompanied by one or more thresholds that trigger a staffing variance beyond the baseline staffing level. For example, one possible infrastructure complexity parameter is the current average (or worst) facility condition assessment (FCA) grade (A-F). It stands to reason that Facilities Management (Engineering) workload is greater in facilities with a lower FCA grade. However, the relationship between workload and FCA grade may not be linear. Perhaps baseline staffing is adequate for FCA grades of A and B, but substantially increases for FCA grades of C. Thus, FCA grade C could be a threshold for triggering a staffing variance within the facility condition infrastructure complexity parameter.

Similarly, if gross square footage is an infrastructure complexity parameter, there are likely thresholds within this parameter at which Facilities Maintenance (Engineering) workload substantially increases, triggering a staffing variance beyond the baseline. Data about links between current staffing levels and various facility characteristics could be reviewed to help inform expert judgments about where to establish the thresholds.

The committee’s conception of infrastructure complexity parameters has a precedent within VHA in models developed by the Office of Occupational Safety, Health, and GEMS Programs (Dulaney 2019a,b). Its staffing models estimate staffing requirements for each of five different job functions (Safety, Industrial Hygiene, GEMS, Emergency Management, and Administrative Support). Each model is applied at the VAMC level. Each job function has its own parameters and associated thresholds that trigger staffing variances beyond a baseline level, but the same parameters are used to define staffing requirements across all VAMCs. Parameters are labeled as “infrastructure” or “workload.” Across job functions, infrastructure parameters include the number of full-time equivalent employees (FTEEs) at the facility, total square feet, VHA facility complexity level, and multi-division/campuses (yes/no). The first three parameters are part of the staffing model for four of the five job functions, meaning that some of the same parameters are relevant for multiple job functions. “Workload” parameters vary from one job function to the next and include parameters such as the amount of Veteran’s Equitable Resource Allocation Methodology (VERA) research dollars, presence/absence of self-supplied utilities, or presence/absence of an onsite medical waste treatment program. SMEs created thresholds that warranted a staffing variance above (or, occasionally, below) for each infrastructure and workload parameter for each job function. The choice of parameters and their thresholds was determined by VHA SMEs, based on careful consideration of data collected about workload requirements in each job function.

The committee learned that there are many differences across VAMCs that impact the Facilities Management (Engineering) workload. Several are typical of any facility maintenance function, including the amount of building space (building gross square footage) and grounds (managed acres), the age and condition of the facilities and internal systems (e.g., HVAC), how the facility (and specific spaces within it) is used, and the number, size, and complexity of construction projects. Some parameters may be particularly or only relevant for health-care facilities, such as space use intensity or the presence of research laboratories. Finally, some parameters may be particularly or only relevant for VHA facilities, such as the requirement to maintain unused space until it is decommissioned.

Parameters That Could Be Considered in the VHA Engineering Staffing Model

This section briefly describes parameters that could be considered in the VHA Facilities Management (Engineering) Staffing Model. Each was mentioned by at least one source as impacting the Facilities Management (Engineering) workload. Some can be inferred from benchmark reports, such as the one published jointly by the International Facilities Management Association (IFMA) Health Care Council and the American Society of Healthcare Engineering (ASHE) in the 2010 report Operations and Maintenance Benchmarks for Health Care Facilities. Others can be inferred from a facilities management staffing tool available to ASHE members. Still, others were mentioned by the study sponsor (Broskey and Alvarez, 2018) and in a variety of presentations by VHA chief engineers and medical center directors (See Appendix A for the agendas and speakers for all the committee’s workshops and meetings). In addition, the committee heard from military manpower planning experts (Scott McCulloch, U.S. Army Manpower Planning Analysis Agency; Timothy Clary, U.S. Air Force Manpower Analysis Agency); facilities managers at other federal agencies (Scott Robinson, NASA; Daniel Wheeland, National Institutes of Health); a public nonprofit academic health-care system (William Seed, Jackson Health System); private-sector health-care organizations (Don Orndoff, Kaiser Permanente); and private-sector property management organizations (JLL; Grant Thornton; Gordion) (NASEM, 2019d).

VHA, of course, could do its own benchmarking study to determine the major ways in which infrastructure complexity varies across VAMCs. It already has extensive data about its facilities and about its Facilities Management (Engineering) departments (staffing levels, work order data, types and amount of work outsourced). Absent VHA-specific benchmarking, the committee prepared the following list and recognizes that not all of the parameters mentioned by different sources are completely unique. Some are partially or fully redundant with each other, and some capture a type of variance that may apply to very few VAMCs. The committee believes that the staffing

model could include a much smaller subset of them and still produce accurate staffing estimates. Following is a list of infrastructure complexity parameters, listed in no particular order:

1. Building gross square feet (GSF)—Every source and presenter cited this as a core driver of Facilities Management (Engineering) workload, and thus staffing requirements. Some property management companies rely primarily on this parameter to establish initial staffing levels, which are refined as the property management company gains greater familiarity with the realities of the property to be managed. It is also more likely to be an overriding parameter when the facilities to be maintained are of similar age and condition across locations, as is the case for some private-sector health-care organizations.
2. Managed acres—This parameter reflects the number of acres that must be actively managed at any given site and obviously impacts workload for grounds maintenance. It varies substantially across VAMCs.
3. Topography and land use at the site—This parameter reflects the fact that grounds maintenance workload varies depending on the type of grounds to be maintained—for example, woodland versus lawns versus flower beds. These differ substantially across VAMCs.
4. Usage of the space to be maintained, or space utilization intensity—This parameter reflects the fact that the Facilities Management (Engineering) workload depends heavily on how the space to be maintained is used. An operating room or emergency care ward requires enough staff to provide frequent preventive maintenance and to respond with surge capabilities if a problem occurs. These types of space also place stringent requirements on when and how maintenance activities can be conducted, which can have ripple effects on other facility maintenance requirements. In contrast, the workload associated with maintaining a parking garage is much lower. This parameter can be categorized in a variety of ways, as the committee heard from several different presenters. The VHA Capital Asset Inventory (CAI) database contains square footage organized down to the floor and department level, so it seems possible to estimate workload according to major types of usage. The ease of doing this may depend on how consistently space is labeled and defined within the CAI database. The VHA Facility Complexity Model incorporates types of clinical services that could also provide a guide for categorizing building space usage. Alternatively, SMEs could create meaningful, broader categories that can be defined so that others understand them, using categories such as ambulatory, operating suites, inpatient, ancillary, administration, laboratory, and parking garage. The committee notes that it may be tempting to use the existing VHA Facility Complexity Model to define space usage requirements. Recall that this model comprises (a) patient population measures, (b) clinical services complexity, and (c) education and research measures. Some of its parameters are undoubtedly correlated with facility maintenance requirements, but the Facility Complexity Model does not capture the full story for infrastructure complexity, particularly for VAMCs in the lower categories of the Facility Complexity Model. For example, a medical center in the lowest category (Level 3) of the Facility Complexity Model may still have a high degree of infrastructure complexity.
5. Building age and condition—On top of space usage, building age varies both across and within VAMCs. The greater the concentration of older buildings, the greater the potential workload for the Facilities Management (Engineering) department. As noted above, even if some of the buildings are no longer used, they must still be maintained to a reasonable degree. The facility condition assessment (FCA) is one reasonable source of information about building condition that is already available. In their presentation to the committee, Walker and Dilinger (2019) argued that it is wiser to focus on the age of systems and equipment within facilities, rather than the age and condition of the building in which they are located. For example, an old building might have a new HVAC system. Walker and Dillinger also identified 72 different systems at Stanford Health Care facilities. VHA Facilities Management (Engineering) currently tracks maintenance on critical systems but does not have a record of every system and piece of equipment in VHA facilities. Therefore, achieving this level of detail about systems within buildings would require a large, dedicated effort to create and maintain, but it would provide a very comprehensive understanding of all maintenance requirements.

6. Buildings on the Historical Register—In addition to building age, some buildings have been designated as historical. Maintenance of these buildings must meet federal standards for such buildings. To the extent that some VAMCs have more historical buildings than others, the Facilities Management (Engineering) workload will vary.
7. Proximity of facilities—Most VAMCs have multiple buildings and some have more than one campus, with multiple buildings per campus, such as the VA Portland Health Care System (VAPORHCS), which has campuses in Portland, Oregon, and Vancouver, Washington, as well as 10 outpatient clinics across central and northwest Oregon. Several VAMCs are now consolidations of several former independent VAMCs, such as VAMC at Dallas, Texas, which now includes Bonham and Kerrville, Texas. In some instances, the campuses are many miles apart, which means that the Facilities Management (Engineering) department needs some staff at each campus and that staff who work at both campuses likely spend a substantial amount of time traveling between campuses.
8. Distance or time spent traveling to remote work sites—This parameter is related to the previous one but can vary substantially across VAMCs, so may warrant separate consideration.
9. Deferred maintenance—This parameter reflects the extent to which planned maintenance has been deferred. It may be further broken out by the criticality of the system(s) for which maintenance has been deferred, the amount of time that maintenance has been deferred, the size or cost of the maintenance activities, and so on.
10. Location of maintained equipment—In any given facility, the equipment to be maintained may be located above, adjacent to, or below the space for which it plays a critical role. This has implications for Facilities Management (Engineering) because it impacts when and how the equipment can be accessed, and whether special procedures (e.g., infection control protocols) must be followed when accessing the equipment. For example, if critical equipment is located adjacent to an operating suite, gaining access to perform maintenance will be less disruptive than if the equipment is located above the operating suite.
11. Weather conditions—This parameter generally reflects the extent to which a given location experiences weather conditions that increase the Facilities Management (Engineering) workload, such as the average number of days requiring snow or ice removal from sidewalks and roads.
12. Climate zone—This parameter serves as an indicator of heating and air conditioning use, which impacts workload associated with operating and maintaining heating and cooling equipment. Indirectly, it also impacts the workload associated with grounds maintenance and fleet management and may impact staffing requirements for construction projects.
13. Ease of work order tracking—This parameter reflects differences in the amount of workload required to track work order completion. Some VAMCs may use an automated work order tracking system, coupled with well-trained staff who use it consistently and accurately. Others may not.
14. Anticipated amount and complexity of the Strategic Capital Investment Planning (SCIP) process for a given VAMC—Facilities Management (Engineering) workload is greater when there is more of this type of planning to do, or the planning process for expected projects is more complex. (Note that SCIP is defined and covered in Appendix C.)
15. Number, size, and complexity of nonrecurring maintenance (NRM) construction projects currently under way—Project engineers must plan and monitor NRM construction projects because their workload is impacted by the number, size, and complexity of those projects.
16. Number of construction projects designed in house—Workload for project engineers is heavier when they do the design work, rather than overseeing contractors who do the design work.
17. FCA dollar value—This parameter is an indicator of the scope of infrastructure projects under way. Higher values lead to more workload for project engineers.
18. SCIP dollar value to correct all gaps—The larger the SCIP dollar value, the more construction work there is to be completed, which impacts workload for project engineers.
19. Dollars of construction—Project management workload is directly related to dollars of construction. VHA project engineers must inspect, coordinate, and oversee all the construction activity, and the higher the dollars of construction, the larger the workload, especially but not exclusively for the CP function.

20. Hospital complexity level where construction projects are performed—Construction projects in facilities or areas providing tertiary care are generally more difficult than construction projects in facilities or areas that provide secondary or primary care, which impacts workload for the CP function.
21. Dollar value of medical equipment replacement projects—Generally, the higher the dollar value of medical equipment to be replaced in a construction project, the greater the workload for the CP function.
22. Rural or urban setting for construction work—Facilities located in an urban setting typically have more complex issues related to construction work (e.g., contractor parking, staging) than those located in a rural setting, which can impact workload for project engineers.
23. Building height—Construction work is more complex for multistory buildings than for buildings with only one or two levels, which can increase the workload for project engineers.
24. Ability to outsource Facilities Management (Engineering) activities—Some Facilities Management (Engineering) activities can be outsourced, such as elevator maintenance, specialized electrical testing, major repair of chillers/boilers, building automation control troubleshooting, fire alarm system testing, and so on. This can reduce the number of FTEs required to perform the workload for some positions or roles (e.g., locksmith, painter), but can also increase the workload for administrative or supervisory personnel.
25. Unique systems/processes—These are systems or processes that fall under the purview of the chief engineer in only some locations, such as a fire department, onsite wastewater treatment plan, or onsite medical waste disposal. When present and under the purview of the chief engineer, these systems and processes obviously require additional staff.
26. Number of functions reporting to chief engineer—The workload for the Engineering Administration function is impacted by the number of different functions that the chief engineer (and leadership staff) must lead. While the sponsor requested a staffing methodology for only three functions, the committee heard that the chief engineer’s office is typically responsible for at least one to two additional functions, including Biomedical and Occupational Safety, and sometimes for functions such as Housekeeping and Janitorial.
27. Unique disaster/preparedness requirements—Some sites may have unique staffing requirements to handle known risks for natural disasters such as being in an area with greater risk of earthquakes or flooding or other extreme weather events.

Even with a list of 27 potential infrastructure complexity parameters, the committee acknowledges that there could be other VHA-specific parameters that are not included in the list. Nevertheless, the list does cover key documented ways in which site-specific parameters cause workload to vary (NASEM, 2019a,b,f).

Number of Infrastructure Complexity Parameters to Include in the VHA Facilities Management (Engineering) Staffing Model

A key decision that VHA will face when it starts to apply the Engineering Staffing Model is the number of infrastructure complexity parameters to include, which will also depend critically on determining the ones for which VHA has or can develop high-quality data. There may be choice points in the initial instantiation of the model where it is better to move ahead with a parameter for which VHA has high-quality data, even though another correlated parameter would be the ideal choice if high-quality data were available.

Statistical modeling techniques, such as regression analysis, can be used to evaluate the extent to which indicators of infrastructure complexity (e.g., gross square footage, facility condition assessments) account for variability across VAMCs in current Engineering staffing levels. The results can provide useful insights to inform decisions about which parameters to include in the modeling effort and where the thresholds for trigging a staffing variance might lie. For example, analysts will be able to identify parameters that account for the most variability in current Engineering staffing levels, the order in which the parameters enter the model (indicating how much additional variance each one accounts for after all others have been entered in the model), and the point at which adding more parameters yield little useful additional value in explaining variability across VAMCs in current Engineering

staffing levels. The committee heard that a reasonable heuristic for determining the number of parameters in a model is that point at which about 80 percent of the variability in current staffing levels can be explained (Norman, 2019). Furthermore, using data from prior years could possibly serve as a way to check the consistency of the analysis results over time. Also, regression analysis can determine which indicators are intercorrelated, allowing the modeling team to reduce the list of inputs while helping ensure that correlated predictors are not, or rarely, employed (see Chapter 3).

Note that these types of analyses can be applied to the staffing of different functions (e.g., O&M, EA, and CP) or at a more global level (e.g., all Engineering staff at a VAMC). To carry out these analyses, it will be necessary for VHA to identify accurate, reliable sources of the input data (indicators of infrastructure complexity) and the criterion that is being modeled, that is, staffing levels. One source for current Engineering staffing levels is the CAPRES database described in Appendix C, but the committee also heard concerns about the consistency and accuracy of this information. Ms. Eileen Moran, Director, VHA Office of Productivity, Efficiency, and Staffing (OPES), described doing this type of analysis to develop an administrative staffing model for VHA (Moran, 2019). She indicated that OPES has also conducted preliminary statistical regression analyses focusing on current staffing levels in VHA Facilities Management (Engineering) departments, using variables available in the OPES portfolio. She indicated that further consultation is needed with VHA Facilities Management (Engineering) leaders to identify the most relevant variables to include in future regression analyses, and how to obtain and gauge the quality of data that are currently not part of the OPES portfolio—for example, CAPRES data.

The committee notes that regression analyses are only as accurate and useful as the data on which they are based, so great care must be taken in selecting the input data (see NASEM, 2019d) and the criterion to be predicted. Furthermore, while regression analyses help identify the most explanatory parameters underlying current staffing levels, they do not indicate whether current staffing levels should be adjusted to achieve the desired performance standards. Still, regression analyses can help inform and narrow the list of parameters to include in an efficient staffing model.

FINDING 4.2: VHA already has data that can be used to evaluate the extent to which several different parameters account for variance across VAMCs in Facilities Management (Engineering) current staffing levels, as well as raw information that could be used to establish high-quality indicators of other potential infrastructure complexity parameters.

NOTIONAL APPLICATION OF THE VHA FACILITIES MANAGEMENT (ENGINEERING) STAFFING METHODOLOGY

Estimating the Number of Facilities Management (Engineering) FTEs Needed at the VAMC Level

Figure 4.5 represents a notional application of the Facilities Management (Engineering) Staffing Methodology. The committee developed this to illustrate the outputs of the staffing methodology, assuming that the baseline staff levels have been established and that a set of infrastructure complexity parameters and their associated thresholds for triggering staffing variances have been selected. As noted earlier, decisions about which parameters to include in the staffing model and the thresholds within each parameter that trigger a staffing variance beyond baseline should be informed by a blend of expert judgment and analyses of links between existing staffing levels and KPIs.

In the figure, the baseline staffing level (expressed as FTEs) is listed at the top, followed by five notional infrastructure complexity parameters that the committee believes substantially impact workload for one or more Facilities Management (Engineering) functions. In the notional application of the methodology, the infrastructure complexity parameters are as follows:

1. Department size by space usage—This parameter combines two of the infrastructure complexity parameters described above: building square footage and space usage. It is based on the amount of square footage dedicated to each of several types of usage that vary in their intensity with regard to Facilities Management

(Engineering) workload, such as ambulatory, inpatient, operating room (OR), administrative, ancillary, and parking garage. It incorporates the reality that Facilities Management (Engineering) workload is higher in some types of space than others and allows for staffing variances above baseline depending on the amount of square footage dedicated to each space type.

2. Facility condition index—This parameter is a way to operationalize the building condition parameter mentioned above. It reflects the fact that the Facilities Management (Engineering) workload is larger when a larger proportion of the buildings at a given location are in poorer condition.
3. Average facility age—This parameter was also mentioned in the earlier list. It may be correlated with the facility condition index just listed, but it also reflects the fact that it generally takes more effort to maintain older buildings, even if the facility condition index is at generally accepted levels. It is also generally true that VAMCs with the oldest buildings will require updating and renovation sooner than VAMCs with the newest buildings, with implications for the Capital Projects workload.
4. Managed acres—This parameter reflects the fact that there will be more grounds maintenance at VAMCs that have more acres to groom and maintain.
5. Planned construction in dollars—This parameter reflects the amount of construction occurring at a given VAMC, which impacts the workload of the Capital Projects function in particular but may also impact the Engineering Administration and O&M functions.
6. Unique requirements—This parameter indicates the impact on workload experienced by Facilities Management (Engineering) departments in VAMCs that have responsibilities or infrastructure or systems that other VAMCs do not, such as a 24/7 staffing for a water purification plant or a fire station.

Each of the first five infrastructure complexity parameters in the notional model has three thresholds, labeled A, B, and C. Threshold A is for the highest level of infrastructure complexity, B for the middle level, and C for the lowest level. These labels reflect levels of complexity in a manner similar to the VHA’s existing Facility Complexity Model, which should facilitate interpretation, as medical center directors are already accustomed to thinking of

Level A as the highest level of clinical complexity. For example, for the square footage by usage type parameter, the Level A variance threshold for ambulatory space is 400,000 square feet, the threshold for Level B is 200,000 square feet, and the threshold for Level C is 100,000 square feet. For the facility condition index infrastructure complexity parameter, Level A is associated with greater than 25 percent of buildings or departments receiving a grade of D or F, Level B with 10 percent receiving a grade of D or F, and Level C with 5 percent receiving a grade of D or F. The final infrastructure complexity parameter, unique requirements, is depicted with one “threshold”: Present. It is possible that some unique requirements could have more than two thresholds.

For each threshold level, the notional methodology output shows a staffing variance above baseline for each of the three functions: EA, CP, and OM. Thus, it reflects the reality that infrastructure complexity parameters impact workload differentially, depending on the function. In other words, the size of the staffing variance can be different in each of the three functions. In some instances, a particular parameter or threshold may not warrant a staffing variance beyond the baseline for a particular function, in which case the assigned variance is zero.

Following is an example of how to interpret the values in the notional application of the methodology (see Figure 4.6). For a facility that surpasses the level A threshold for ambulatory square footage (more than 400,000 square feet), the figure shows that 0 additional FTEs beyond baseline are required in the EA function, 3 additional FTEs are required in the OM function, and 0.75 additional FTEs is required in the CP function. For a facility that surpasses the level A threshold for operating suite square footage (more than 112,000 square feet), 0.5 additional FTEs beyond baseline is required in the EA function, 3 additional FTEs are required in the OM function, and 1.25 additional FTEs are required in the CP function. Note that only one staffing variance is applied for each parameter (or subparameter within department size by usage) threshold. In other words, if a VAMC meets threshold A for a given parameter, then the staffing variances for threshold A are added to the baseline staffing level, and the staffing variances for threshold B and threshold C are ignored for that parameter. Similarly, if a VAMC meets threshold B

for a given parameter, the staffing variances for threshold B are added to the baseline, and the staffing variances for thresholds A and C are ignored for that parameter.

Each VAMC can be profiled according to the infrastructure complexity parameters and associated threshold levels. Once the VAMC profile is established, the staffing values in the columns can be summed to create a total number of FTEs for each of the three functions. Figure 4.6 provides an example of a notional VAMC. In the figure, staffing variances were retained for each threshold that the VAMC meets. This makes it possible to sum the baseline staffing and complexity-related variances to calculate the total estimated FTEs for each of the three functions. As noted above, the application of the staffing methodology must also incorporate adjustments to the FTE levels to reflect worker non-available time due to holidays, rest breaks, and leave.

In Figure 4.6, the notional VAMC meets threshold A for square footage dedicated to ambulatory, operating suite, inpatient, and ancillary purposes, and threshold B for square footage dedicated to administrative and parking garage purposes. The notional VAMC meets threshold C for the facility condition index, threshold B for average facility age, threshold A for managed acres, and threshold B for planned construction dollars. Last, this VAMC has a water purification plant and a noncontiguous campus. The model output suggests that this VAMC needs 2.25 FTEs beyond the baseline level of 7 FTEs in the EA function, 22.75 FTEs beyond the baseline level of 20 FTEs in the OM function, and 8 FTEs beyond the baseline level of 6 FTEs in the CP function. Across all three functions, the model output suggests that this VAMC needs 66 Facilities Management (Engineering) FTEs, keeping in mind that this figure does not include occupational safety and health or biomedical staff that might also fall under the purview of the chief engineer at this location.

The notional model application shown in Figure 4.5 does not drill down to the level of specific jobs within each of the three major functions, but the same process could be used to do so, as illustrated in Figure 4.7. Alternatively, the model could be applied at the level of the three functions, leaving it up to the chief engineer to determine how best to allocate the FTEs across jobs in a given VAMC. Ultimately, each Chief Engineer will also need to make recommendations about which jobs or types of work can or should be outsourced. Contracted labor can be an important tool in any staffing model. The staffing methodology proposed in this chapter assumes that the workload requirement is the same, regardless of whether the work will be performed by a government employee or a contractor.

Estimating a Staffing Budget Request at the VAMC Level

To build a staffing budget request, the committee envisions that the chief engineer at each VAMC will apply known salaries associated with the number and type of FTE positions to be filled and/or known or estimated costs to outsource some types of work. The committee heard one presentation in which a university-based health-care system was able to derive detailed costs associated with various activities performed by Facilities Management (Engineering) staff (e.g., Walker and Dillinger, June 11, 2019). If this level of cost information were available, it could be used to explore “what-if” scenarios about how changes in staffing levels could impact costs or, conversely, how investments in new technology or updated equipment might drive down costs for labor. The staffing methodology described in this chapter can be used to do rough “what-if” analyses—for example, discussing the number or scope of additional NRM projects that could be undertaken with more staff, and the relative offset in FTEs required to maintain facilities in their existing state but would require discussions with the medical center director about how exceptions to the model output differ from impact to the staffing variances associated with the already-defined infrastructure complexity parameters, thresholds, and staffing variances above baseline.

SUMMARY

In summary, this chapter describes a staffing methodology that accounts for the fact that Facilities Management (Engineering) workload varies across sites due to differences in infrastructure complexity. It begins with a baseline that is common to all VAMCs and builds from the baseline in a logical way by associating staffing variances from baseline with defined infrastructure complexity parameters and thresholds. Once the model is created, a chief engineer at any VAMC can explain how the staffing estimates (model output) were derived in a way that can be readily grasped by the medical center director and other stakeholders. The VHA Engineering Staffing Model is scalable and adjustable as conditions change. It focuses on the VAMC level because that is where decisions about the budget allocation for Facilities Management (Engineering) staffing are made. As with any staffing model, the devil will be in the details.

The notional application is intended for illustrative purposes only. The committee is not dictating that these, or only these, parameters should be adopted, or that the thresholds should be defined as shown in this chapter, or that the staffing variances are accurate. The point is to illustrate how the application of the methodology leads to FTE outputs. The input required to build a model is based on estimates of workload for performing different functions, the thresholds that trigger a necessary staffing variance beyond the baseline, and the size of the variance associated with each threshold.

Absent clear performance-level expectations for Facilities Management (Engineering), SMEs will need to adopt some shared understanding of what they believe the performance levels should be so that they can move forward with determining how much workload is required to achieve those levels. Perhaps the initial “understanding” will need to be “whatever level of performance is being achieved right now” until policy can be developed and promulgated to better inform SME judgments.

The committee acknowledges that it will require significant effort and input from VHA personnel to identify, define, and justify the choice of infrastructure complexity parameters, thresholds, and staffing variances (in FTEs)

trigged by each threshold. Chapter 5 describes a process for compiling the necessary information and expert judgments. It emphasizes the importance of collecting input from various types of experts, including reality checks by “operators” (nurses, medical staff, technicians), and gaining buy-in from VAMC and VISN-level leadership. No staffing model can be perfect from the outset, so it will be important to engage in a process of continuous improvement. Over time, infrastructure complexity parameters that are not viable now because the necessary data do not exist may become viable as new or higher quality data become available. Over time, as the context in which the model operates changes, the infrastructure complexity parameters, thresholds, and staffing variances may need to change. The point is that the methodology described in this chapter can change in transparent ways, which allows for discussion and clear understanding of how Facilities Management (Engineering) staffing estimates are being derived, just as the VHA Nurse and Occupational Safety staffing models do. The VHA Facilities Management (Engineering) Staffing Methodology provides staffing estimates based on objective data that can be extracted from VHA systems, coupled with informed judgments by individuals who understand the type and amount of work to be performed to ensure that facilities are continually ready to “honor America’s veterans by providing exceptional health care that improves their health and well-being.”²

___________________

² VHA Mission - Honor America’s Veterans by providing exceptional health care that improves their health and well-being. See https://www.va.gov/health/aboutvha.asp, accessed December 2019.