Proceedings of a Workshop

Facilities Staffing Requirements for the Veterans Health Administration—Operations and Maintenance of the Physical Plant and Equipment

Proceedings of a Workshop—in Brief

The National Academies of Sciences, Engineering, and Medicine convened a workshop on February 5-6, 2019, to explore the responsibilities, challenges, industry standards, and best practices associated with the staffing required to operate and maintain complex and diverse physical plants, as in the Veterans Health Administration (VHA): medical centers, hospitals, outpatient clinics, skilled nursing facilities, and support facilities. This workshop was part of a larger effort, undertaken by an ad hoc committee of the National Academies for the Department of Veterans Affairs (VA), to prepare a resource planning and staffing methodology guidebook for VHA Facility Management (Engineering) Programs.

During the workshop, facilities engineering experts from government, the private sector, and academia met with several committee members. Presentations by individuals and panelists along with discussions among all meeting participants covered many topics, such as impacts of plant operations and maintenance (O&M) on patient care and outcomes; measuring O&M effectiveness; cost management; regulatory and accreditation requirements impacting plant O&M; O&M staff competencies; and customer-service feedback. No conclusions or recommendations were drawn from this workshop; nevertheless, the large amount of information gained here plus information anticipated from future workshops will significantly inform the more comprehensive committee effort.

Workshop Co-Chair Colin Drury, SUNY Distinguished Professor Emeritus, University at Buffalo, opened the first day by welcoming attendees. He then explained more about the larger committee effort, which includes identifying key variables that must be factored into the resourcing methodology and making recommendations for staffing models and transitioning from current staffing strategies. He noted that the VHA has 18 regions and 172 sites, which vary in age (modern to historic), complexity (different levels of medical treatment), location (urban to remote), and size (large to small). A chief engineer at each site oversees workers, with the aim of ensuring that the facilities remain operable and meet all requirements for medical staff and patients. Drury finished by describing the larger committee and its need to acquire a large amount of data about the VHA—such as how it performs O&M; details of required knowledge, skills, and abilities (KSAs); and how O&M modeling works—in this and future workshops.

CONTRACTING APPROACHES TO ENHANCE VHA FACILITY O&M SERVICES

Co-Chair Drury introduced the keynote speaker, Mark Tomassoni, KBRwyle, who began by framing the presentation. First, he explained that the VHA currently does not outsource total facility O&M services, but it does outsource smaller O&M services (landscaping, custodial), major new engineering and con-

struction, as well as minor repair and upgrades. Second, he offered the following list of recommendations for the larger committee:

1. Compare the costs of providing total facility O&M services under different types of contracts versus through government civil servants.
2. The Office of Management and Budget (OMB) Circular A-76 is a possible tool to conduct competitive comparisons between government civilians versus contractors.
3. The VHA may be able to conduct competitions to outsource facility O&M.
4. Several test-case facilities should be established for 2-3 years.
5. The VHA could outsource facility O&M at single medical complexes, or contracts could be considered for Level 2+ medical facilities (meaning facilities of mid-to-lower complexity).

Tomassoni supported his conclusions by using Department of Defense (DoD) examples that he suggested could apply to the VHA, such as DoD acquisition guidelines (firm fixed-pricing, performance-based contracting); advantages and disadvantages of OMB A-76 (cost savings but very time consuming); various types of contracts (including the advantages of indefinite delivery/indefinite quantity [IDIQ] multiple awards); common work scopes and types of work (preventative maintenance, service orders, watch-standing for power and steam); small business set-asides versus unrestricted contracts; standard templates used by the Navy for facility-related contracts; and two major types of DoD facility O&M privatization. He closed with the following five issues for committee consideration and decision:

1. Which, if any, existing DoD facility O&M structures could be adopted by VHA.
2. How and where VHA facility O&M acquisition is to be managed.
3. How, or if, portions of VHA facility O&M can be privatized.
4. How, or if, work could be converted using A-76 or another “direct-to-contract” process.
5. Whether the VHA should establish beta test sites as soon possible and run for 2-3 years.

As expected, Tomassoni’s presentation gave rise to much discussion. For example, committee member Robert Anselmi asked about figuring out staffing needed—just using square feet, or using complexity factors, and how detailed? Tomassoni said that they use estimators, just as in the construction or base-operations world, and documented metrics to assess numbers of people needed for different types of services, like mechanics, electricians, vehicle maintainers, groundskeepers, and so on. Also, many trade associations and private companies have that kind of information. But it takes time and energy to distill the information and convert it to something usable, such as a staffing model that develops a simple algorithm to indicate full-time equivalents (FTEs) required for a job. Another point Tomassoni touched on involved protests or lawsuits filed by many companies against government acquisition organizations because of inadequate workload data. If the VHA decides to outsource services, it will need much detailed historic workload data showing facilities, service orders, preventative maintenance events, minor repairs, and so on; that is a major takeaway.

Further examples were (1) in connection with workload data, discussions of a system called Maximo being used by about a third of VHA facilities (replacing the work order system Automated Equipment Management System/Medical Equipment Reporting System [AEMS/MERS]), although Maximo did not meet all needs, and the VHA is looking at an alternative called Defense Medical Logistic Systems (DML) that also has a facilities management package; (2) VHA preferences for service-disabled veteran-owned small businesses (SDVOSBs), which limit the VHA’s ability to contract for outside services; and (3) concerns about losing in-house expertise if the VHA were to go with a privatized O&M solution, which could result in the VHA facing higher bids from outside contractors in the future.

Co-Chair James Smith, U.S. Ambassador (retired), asked how much data is important in order for the VHA to have a manageable model in terms of staffing and resource allocation to 141 sites? Tomassoni

replied that the VHA has collected staffing information across all centers, probably categorized by types of workers, kinds and sizes of structures, geography, and so on. A simple algorithm could be developed using those metrics to produce a basic equation. Start with that, apply some factors (for age, perhaps), and estimate staffing based on various elements, like square footage, and so on. Gary Krupa, VHA representative, noted that the VHA does have facility condition assessment (FCA) programs, which track major systems by age, be it 5, 10, or 50 years in some cases; the VA is in the process of converting that to another program known as DoD BUILDER. John Dodier, chief engineer, Portland, Oregon, VHA Medical Center, explained that BUILDER goes down to the component level, but that the VHA is just starting to use some of the applications.

Tomassoni went on to describe his type of analysis using BUILDER. The BUILDER software works from the bottom up, but when preparing a bid, one needs to start with an order-of-magnitude estimate of what is needed at the top level to provide the services promised. This top-level information can be, for example, square footage, number of people, and so on. One can then use BUILDER to improve the estimates at the top level and close the gap from the bottom (component) level as knowledge improves. Smith later asked, “How do you deal with the risk as you are bidding a contract?” Tomassoni’s short answer was, “Well, that is the $10,000 question, right? Every line item in the contract has risk associated with it.” For risks like elevators, he relies on professionals.

MANAGING FACILITY O&M AT THREE DIFFERENT VHA COMPLEXITY LEVELS

Robert Anselmi next introduced chief engineers from VHA medical centers with complexity Levels 1 and 2.¹ John Dodier spoke first. The Level 1 Portland healthcare complex (Figure 1, left) is large, consisting of (1) two campuses, Portland, Oregon, and Vancouver, Washington, and (2) a group of leased properties, including five large clinics, two smaller clinics, and two contract clinics. The campuses occupy a total of about 1.4 million square feet and 80 acres, with 3,000 parking spaces (many employees use bicycles, carpools, and buses). Dodier’s current staff consists of about 110 FTEs, plus a contract radiation safety officer and 7 compensated work therapy (CWT) employees. Last year, his group handled approximately 41,000 maintenance and repair work orders plus 6,400 biomedical work orders.

Portland has an electric shop, a pipe-plumber-machine shop, a paint-carpenter-locksmith shop, an energy section, a contract for elevator maintenance, and staff for grounds, transportation, and administration. Vancouver, which is assisted by Portland, has a smaller staff. Illustrative attributes of Dodier’s management philosophy are an attitude of continuous improvement and dedication to the VHA. Some example points from his list of expectations and outcomes are as follows: see a problem, make a plan to get it fixed; never knowingly encumber patient-care space without planning; always be compliance ready; use contractors for construction and specialized maintenance (radio towers, elevators); best not to depend on contractors for emergency response; and keep facilities within budget, attractive, clean, and safe.

In closing, Dodier shared lists of his many experiences—the following observations are illustrative: the recruiting process takes 6 months; it is difficult to recruit electricians and plumbers; there is too much recurring training (privacy, security), and required training offers little technical improvement; barriers to productivity are sick leave, building age (a lesser factor), lack of product standardization (a high factor), and quality of supervision, attitude, and culture.

Jacob Yoder, chief engineer, introduced the Level 2 Bath and Canandaigua VA Medical Centers in New York (middle of Figure 1; note the bridge, which requires maintenance). The two campuses have 95 buildings total (1.5 million square feet) and occupy a total area of 373 acres, of which 172 acres are

___________________

maintained. They have four associated outpatient clinics. Both campuses are “historic”—an issue with state historic preservation offices if modifications are necessary. Yoder’s total staff consists of about 180 FTEs; he said when things are broken, they fix them.

Yoder described the shops he oversees—electric, HVAC (heating, ventilation, air conditioning), pipes, paint, carpentry, and lock—plus boiler plant, grounds work, and other engineering staff responsibilities. His mission is keeping the medical center safe, operational, and current with requirements and trends in services supported (such as needed maintenance and repair, renovation and new builds, provision of utilities, possibly some service delivery, and support of other requirements). Yoder’s expectations were that staff accomplish the mission in their areas of responsibility within all applicable guidelines; if unable to do that, staff should notify him immediately for direction. Illustrative of his long list of duties are the following: establish policy, procedure, process; quality oversight; final decisions on all hiring (verify competence); ask the right questions and lead; work with little oversight.

After showing some performance indicators for measuring success and identifying issues, Yoder discussed the AEMS/MERS; although still in use, he indicated that it is not well supported, training is nonexistent, the VA has already moved away from it, and something new is needed. Yoder offered thoughts on modeling, noting that differentiation between in-house staffing versus contracted services is challenging, and he stressing that utilized-maintained square footage is the most important variable (other potential variables include developed-maintained acreage, inventory of systems supported, and facility complexity). In closing, Yoder touched on staffing issues, such as inability to recruit prior to separation, VA hiring preferences and purchasing-contracting processes, and limited classification and pay options; he strongly supported the CWT program (a “best practice”).

The third panelist, Art Ontto, chief engineer, Level 3 Iron Mountain VA Medical Center in Michigan, was delayed by weather and arrived on Day 2. His contribution is included here for completeness. The Iron Mountain facility is our nation’s most rural VA medical center and part of a city of population 7,600. The 27-acre campus comprises 17 buildings, with about 300,000 owned square feet (main building is at right of Figure 1). Seven associated clinics are 80-225 miles away. Ontto’s staff of about 37 FTEs includes a project section, two boiler groups (one for equipment support and the other for boiler operations plus transportation and grounds), and a maintenance and repair group with mechanics, HVAC technicians, electricians, pipe-fitters, locksmith-masons, and carpenter-painters.

Ontto’s 2018 maintenance and repair summary showed about 5,350 preventive work orders and 5,200 repair work orders; that summary also showed, for 14 people, 3,700 hours on sick and annual leave, 382 hours for mandatory training, and 1,232 hours off for holidays. Ontto’s responsibilities, expectations, and measures of success included outcomes like successful accreditation, flawless facility operation, continuous code compliance, proper training of employees, customer-service and patient-satisfaction surveys, results of inspections, and equipment performance. His wishes included, for example, having enough skilled mechanics to have someone on call. Keeping him up at night were worries about conditions such as having no off-hours coverage, something breaking and no one answering the

phone, winter bus trips with patients, complexity of new equipment and systems, and employee workload and burnout.

Among Ontto’s many challenges were taking 10-12 months typically to fill a position (varying from 5-18 months), hiring staff with desired qualifications, needing a work-order system to provide accurate information, spending limits on credit cards for construction and services, lack of contractors in his isolated area, and getting parts and qualified vendors to support breakdowns. To illustrate, he described the harrowing week of January 26-31, when there was difficulty getting help during very cold weather accompanied by a series of breakdowns and alarms. Some general improvements were suggested, such as ability to supplement staff by contractors when needed, complete most preventative maintenance, have a more comprehensive proactive maintenance program, exercise interior valves, and focus on operation of systems versus maintenance and repair.

There was much discussion involving these three presentations. Several examples follow.

1. Committee member Fred Switzer asked how long, when somebody leaves, does it take not just to bring someone new on board, but for that person to get up to a level approximating the previous person’s productivity? Both Yoder and Dodier responded in detail. To sum up, much bureaucratic process (such as position descriptions) and many months are needed to bring a new person on board, and after that it takes more bureaucracy and added time, several months perhaps. (Hiring times for Dodier and Yoder appeared to be somewhat less than Ontto’s typical 10-12 months.) Elaborating, Krupa mentioned that some past managers said, “We’re going to save salary for half a year, so we’re not even going to try to hire for 6 months or 12 months, because I’m going to pocket that money and use it for other purposes.” Dodier replied that in 2007-2008 that was intentional due to poor budgets.
2. Another area of questioning went to differences between facility Levels 1, 2, and 3. Dodier believed that at Levels 2 and 3 (more rural, smaller), everything costs more money than in a big city (Portland is very competitive). Yoder thought that part of what is seen reflects the difference in scale. “Losing one body when you have seven electricians is a lot different from losing one body when you’ve got two or three. There’s no comparison, because the pain you’re going to feel is different, and your ability to cover the most important things is very different.” He added that facility total size matters (e.g., 4,000 folks at Dodier’s versus 500 at Yoder’s); there is a lot less room for shifting workloads when planning a reduction of force at 500.
3. A function-related discussion considered research, which takes place at some facilities but not at others. For instance, Dodier talked about research at his facility and that somehow his engineering elements gain money through research—22 percent of the facility is paid for by research because all the different research programs are 22 percent of the facility footprint. Portland gains FTEs through research, which consists of wet labs, dry labs, and research on animals (mice) and patients (smoking, alcohol). VA representative Steve Broskey added that medical research environments require very specialized HVAC, and pressurized labs have to be maintained very critically. So reliability of the building systems in these areas, the resilience and redundancies, take dedicated attention from the O&M staff—another level of complexity and care to ensure that the facilities in which this research occurs are protected at all times.
4. Ontto responded to a question from Co-Chair Smith about what percentage of an employee’s time is actually documented in the work order effort. Ontto said that they document all their time on what they do, for work orders and preventive maintenance. However, Ontto indicated that the VHA does not require employees to justify their daily 8 hours.
5. Responding to a question by VHA representative Jason Sawyer about average staff time on the job (15-20 years or fewer?), Ontto said he had been there 12 years and hired every person there except two.

6. Anselmi moderated discussions by Dodier, Yoder, and participants concerning (1) written answers to many committee questions about tradespeople close to the actual O&M work and (2) additional verbal questions. On the written materials, both managers agreed with most answers and elaborated where necessary. One example concerned monitoring employees’ work time: Dodier said that this needs to improve to justify the entire work year, not just one-half or two-thirds. Yoder liked the idea of achieving a time target, an expectation for supervisors, but with unions, some supervisors felt it was not worth chasing the time per job unless it was consistently very high versus a standard. Dodier clarified that he was not thinking daily but rather of ensuring that a worker had 1,760 hours of work per year.
7. During the additional questioning, among other topics, a participant asked if there is a growing need to look at cybersecurity and perhaps add more people for it. Dodier said that digital controls are on their own network, but there is a growing awareness to be careful when vendors come with thumb drives, which need security features, and Portland needs security software to protect against viruses. He indicated that they are aware and can probably learn more to better protect their systems, but they are fully separated from the patient-care business network. After more discussion, Krupa responded that cybersecurity has been an area of growing concern. There is a Physical Security Design Manual, which should be updated and released later this year. This manual talks about electronic security and addresses systems security. But there is always tension between the stand-alone systems and interfacing them with others. Krupa mentioned Web-based systems and vendors wanting to dial in, and confirmed, “Yes, it’s an increasingly difficult area that we’re trying to traverse and we’re trying to keep them separated as much as possible, but there are those growing overlaps, especially Web-based systems.” After indicating that several systems (prox card, panic alarm, the VA’s Office of Information and Technology) are on the VA network, Dodier agreed on the need to do cybersecurity better, as “an increasing part of our job.” Yoder agreed as well.

MANAGING FACILITY O&M AT PRIVATE-SECTOR HOSPITALS

Facility managers from two private-sector hospitals made presentations to the committee, the first on Day 1 (with follow-up on Day 2), and the second on Day 2. Both are treated here along with discussions from their panel. Stanley Larson, chair, Division of Facility Operations, Mayo Clinic, Rochester, Minnesota, first summarized Mayo’s strategic plan, with its primary value: “the needs of the patient come first.” After discussing the flow-down from Mayo’s values-mission-vision statements to many responsibilities under facilities, Larson described his key organization elements—two section heads for facilities operations (one for clinical and business, the other for hospitals); a third for waste management, recycling, and environmental services; and a fourth for energy management, utilities, and infrastructure systems. A fifth unit (same level as section heads) manages procurement for facilities operations. At Rochester, there are approximately 13.6 million square feet—15 percent for research and 85 percent spread almost evenly among three other categories: hospital, ambulatory, and business. Other nationwide Mayo units, discussed later, occupy about 10.2 million square feet.

Larson’s staff is large, consisting of 1,050 FTEs, approximately half belonging to unions, and they are responsible for many diverse systems—mechanical (HVACs, much refrigeration, thousands of gas inlets and outlets, water, sewer, pumping, etc.); electrical and electronic (power distribution, lighting, thousands of cameras and card readers, alarms, patient beds and lifts, etc.); and building infrastructure (paint, carpentry, locksmiths, grounds). Campus monitoring is from a facilities operations control center, which operates and can dispatch maintenance staff and respond to alarms 24/7 (there are 1.5 million alarm points). Annual work orders total about 400,000 (220,000 corrective, 180,000 planned). He described the two-campus steam-generation, chilled-water supply, and peak-power loads needed for buildings, surgery, and data-center equipment. Larson displayed a detailed spreadsheet showing

the great range of services that had to be staffed as well as the FTEs and KSAs required at different locations to handle the multitude of O&M tasks and responsibilities. Also, he began discussing a series of spreadsheets showing square footage and FTEs across the Mayo enterprise, a way of leading into his Day 2 discussions of possible modeling approaches using the metric square feet per FTE.

On Day 2, Mark Bricker, senior director, Facilities Service, Children’s Hospital of Philadelphia, began by noting that his hospital first opened its doors in 1855; now there are two modern hospital buildings occupying approximately 1.8 million square feet, plus additional buildings for ambulatory patients, medical research, and business occupying about 5.1 million square feet. The main campus, which is packed densely in an urban area, consists of a main hospital, an ambulatory center, two research buildings, and two additional buildings. He said the main campus is served by self-generated and contracted utilities (electric power, boiler plants, and chillers). Bricker’s facilities services consist of two primary teams: building systems (50 FTEs for operations, refrigeration, and ventilation) and maintenance (76 FTEs for electrical, plumbing, painting, and carpentry, plus operating a service center).

Around-the-clock shifts are obviously necessary for supporting a medical facility. To illustrate benchmarking, Bricker showed data relating his current usage of FTEs to cover various types of square footage to 2013 International Facility Management Association (IFMA) benchmarks—for example, HVAC and Central Plant service: ~365,000 square feet/FTEs (3 shifts, 7 days/week) versus benchmark: 283,000 square feet/FTEs (1.7 shifts, 5.5 days/week). He also showed cost data per square foot covered by his teams. High reliability of systems is important as well, and Bricker showed data demonstrating that (goals are green for >98 percent up-time, yellow for 97-98 percent, and red for <97 percent). For service calls, he showed data indicating about 92,000 calls per year, of which ~83 percent are immediate. Bricker closed by expressing pride in the number of room-upgrade completions his staff was able to complete since May 2018.

After Bricker’s presentation, Stanley Larson went through the remainder of his prior presentation, this time focusing on possible ways that allocations of FTEs could be analyzed on a facility-by-facility basis. His analyses also illustrated potential modeling approaches for the committee’s future work. In summary, Larson displayed a series of spreadsheets, starting with the Rochester location, showing the metric square feet/FTEs for each category (business, hospital, ambulatory, and research). Then, by looking at square footage and FTEs at the other Mayo locations and comparing against Rochester, he could assess if they were roughly in line with Rochester’s data or if their current staffing was perhaps too high or too low. Clearly, economies of scale and site-to-site details need to be considered here, but Larson offered the data and methodology, for comparison purposes. He also showed data for many different maintenance operations that revealed a wide range of staff-to-supervisor ratios (6:1 to 20:1), and he included spreadsheet data showing costs per square foot.

Illustrative parts of the discussions relating to these Day 2 presentations are summarized here. Concerning work orders, Mike Crowley, Jensen Hughes, asked Bricker what kind of metrics he uses for completion of a time frame. Bricker explained that they have immediate, same day, or next day for urgent requests, and then 4 to 7 days for routine requests. He added that they have not been doing well with routine requests, some averaging 10 days, but that is acceptable for a routine request. Co-Chair Drury commented to Larson that his discussion of allocating FTEs was based on two things: gross square feet and the type or usage of those square feet (business, ambulatory, etc.), so there are now two factors instead of one. Drury went on to say that he thoroughly agreed that economies of scale are important here because in a small facility one may well still need to have a boiler-room staff and other essentials. Thus, in a small facility, one cannot go below a single headcount in any of the essential categories; things flatten out at the bottom, and there is not a linear function right down to a zero headcount. Larson agreed.

ADDITIONAL PERSPECTIVES ON MANAGING FACILITY O&M

Co-Chair Smith introduced three additional presentations on Day 2 offering valuable perspectives on fire protection, facility optimization techniques, and facilities management in general. The keynote was delivered by Crowley on NFPA 99 (a National Fire Protection Association code for healthcare facilities). Crowley began by explaining that prior to 2012, NFPA 99 was a standard related to occupancy; however, because risk to patients comes from procedures, not occupancy, in 2012, NFPA 99 chapters were transitioned to a risk-based approach. He then showed an overview of the document’s 16 chapters, which covered standards relevant to fire protection of equipment in hospitals, such as electrical systems and equipment, HVACs, and gas systems and equipment. Chapters also addressed management of emergencies and security. Crowley noted major changes in 2018, such as clarifying responsibilities for risk assessment and a new chapter on dental gas and vacuum systems.

Some key early points by Crowley were (1) NFPA 99 is intended for persons involved in design, construction, inspection, and operation of healthcare facilities, for managers of VA facilities, maintenance managers and supervisors, and officers who ensure that VA facilities comply with NFPA 99; (2) the document provides minimum requirements and applies to all healthcare facilities (except for home and veterinary); (3) construction and equipment requirements apply only to new construction (i.e., completed after September 6, 2017), but also to alterations, renovations, or modernization of existing VA facilities (i.e., those completed prior to September 6, 2017); and (4) inspection, testing, and maintenance requirements apply to both new and existing facilities. Crowley went on to describe risk categories (major injury or death, minor injury, discomfort, no physical impact) and risk assessment, specific chapters relating to equipment and systems, and linkages to other NFPA documents.

After Crowley’s briefing, an interesting exchange started with Anselmi asking about information on metrics, such as times to perform maintenance on the systems described, and ended with a discussion of performance-based maintenance that involved Co-Chair Drury, Bricker, and Crowley. First, Crowley said to Anselmi, “That is a difficult question to answer. All I can say is you can tell a difference between a facility that does the job right and one that does not. It is almost obvious when you walk in. And the manpower it takes to do it. That is a good question. … We have always looked at focusing on what is the biggest bang for the bucks and then justifying from that standpoint.” Crowley went on to say that the manpower demand for maintenance had been reduced over the years. Then he added, “The next step is going to be a little tougher. One of the things we have worked with some of our DoD facilities is on the performance-based inspection, testing, and maintenance. . . . I am not sure I can answer your question with an actual number of people, but I think there are some programs you could look at that could help you reduce those things because, again, the VA is big enough that you can set your program up like we did at the DoD and then go to the authorities you have in your jurisdiction. . . . And then through your division and your safety people probably set a modified program to reduce that demand. Again, that may be kind of bold and people might not like the way it looks in the press if it comes out. It is a valid method, and actually the DoD has been doing it for almost 20 years now on a lot of their Air Force bases and other facilities.”

Co-Chair Drury said that is done in aviation to prevent crashes, an example being knowledge that cracks grow at a certain rate, they can be inspected for size, and it can be worked out how often they must be inspected to be x percent certain of finding them before they become dangerous. Returning to healthcare, Bricker talked about hospitals, ending with, “Currently, in health care, we are not allowed to do performance-based maintenance on facilities’ equipment.” Crowley’s last comment was directed at the pass rate, or how assured one can be that something will not fail: “You have to select your pass rate. Is it 95 percent? Is it 99 percent assurance rate? That is the biggest challenge we have. Where do you set that? You are a hospital. Do you set it at 95 or 99 or 99.99? Where do you set that acceptance rate for failure? That is the big question.”

Anselmi introduced the last two panelists, who discussed management of a central utility plant (CUP) and overall facilities management. The first speaker was Daniel Wheeland, director, Office of Research Facilities, National Institutes of Health (NIH); he described data-driven operational excellence at the Bethesda, Maryland, Campus CUP. That 300-acre campus has 75 buildings (~12 million square feet), for which CUP reliability is crucial. After a 2014 reorganization, CUP was maintained by a new division, and building maintenance was led by another division, an arrangement requiring both divisions to work well with one another. Wheeland believed pride, performance, professionalism, and accountability are enhanced by aligning workforce to real property assets. To manage this CUP, one of the largest under a single roof, reporting to Wheeland are a chief engineer, a chief architect, seven divisions (planning, O&M, technical resources, environmental protection, budget and financial management, design and construction management, and facilities stewardship), and two offices (acquisitions and hospital physical environment). Energy inputs of fuel oil, natural gas, and electricity are converted to huge amounts of chilled water, steam, electricity, and compressed air flowing through a utility distribution system that services the NIH buildings.

Wheeland focused on data as the foundation for operating safely, reliably, efficiently, and cost effectively (“You can’t manage what you can’t measure!”). The CUP data platform currently collects over 34 million data points per day, leading a paradigm shift that includes automatic fault detection, predictive maintenance, documented standard operating procedures for O&M, load forecasting, and optimization; the result is executive-level daily performance data accessible with one click. The data platform powers a multidisplay CUP control room that enables operational intelligence like fault detection, root-cause analysis, shift performance, and reliability-centered maintenance. Wheeland discussed digitization and optimization in more detail, leading to the key outcomes of increased efficiency (e.g., run minimum number of chillers, shift load to most efficient chillers), enhanced reliability, and tangible reduced cost—millions of dollars saved despite increased utility costs and demand.

Juan Ontiveros, associate vice president, Utilities, Energy, and Facilities Management, University of Texas, Austin, spoke next. He described the large area occupied by the university and the energy generation and distribution, which serves the campus, as a backdrop to his major theme—a “holistic approach to total energy.” To illustrate the power of this approach, Ontiveros presented a graph of the effects of utility improvements over the years 1976 to 2018 that showed a doubling of campus space (from about 10 million to 20 million square feet) but virtually the same number of BTUs in 2018 as in 1976, with a return to 1976 levels of carbon emissions and fuel. To achieve this performance, he audited the operation and developed an action plan. Equipment (especially remaining useful life) was assessed and identifiable immediate risks were prioritized, along with operational opportunities, organizational structure, and personnel skill levels.

Ontiveros’s organizational objective was creation of a self-sustaining process, which involved a layered organization, optimized hiring, and training that takes advantage of the layers. The layers themselves were set up to permit both horizontal and vertical movement as people grew in their jobs (operators moving from trainee to higher levels, assistant supervisors to supervisors, managers to associate directors). Hiring was aimed at keeping employees within the department, helping morale, and loyalty. Training was designed and developed, along with training materials and implementation procedures (e.g., a craftsperson was given 18 months to complete additional training for advancement; employees were paid up to 6 hours overtime per week to do training modules after hours; once certified for a position, employees can precertify for another [no extra pay]). Pins, photos, and framed certificates added incentives. In conclusion, Ontiveros reemphasized the holistic approach, creation of a master plan, involvement of all players, recognition that changing and improving systems involves smart risk-taking, and the need to be methodical.

A couple of discussions after the presentations are illustrated here. Smith asked Wheeland if operators for boilers, chillers, and so on also run the control system, or whether there are separate people managing it. Wheeland replied that the organization used to consist almost exclusively of wage grade,

wage leader, wage supervisor, and maybe a GS-13 in charge, who had worked his or her way up. However, they now have a more diverse workforce, so controls people are specialists, whereas blue-collar operators are doing some of the same functions they used to do, but in a more informed and less labor-intensive fashion, and over time the ratio of blue collar to white collar is evolving more in the direction of white collar. Wheeland added that sometimes there is trepidation on the part of the people who grew up operating in a legacy environment, and they do not like differential equations, machine learning, and artificial intelligence. For them, it has been quite a change-management challenge, one that was originally met with a lot of fear. Summarizing, he believed that there has been positive developments, much change in management, and a lot of trust building and confidence building that blue-collar workers are really important, and as the skills of both blue-collar and white-collar workers were harmonized and incentives were created for successful performance, some of the people who were against this change are now satisfied.

Oleh Kowalskyj, VA representative, asked Ontiveros about salaries paid in order to be competitive. Ontiveros confirmed that a supervisor receives approximately $100,000, an entry level about $48,000, the next level about $60,000, and the next level above that about $70,000. But he said it is very hard to find those people; one does not want to lose them, because they are an investment. So their goal is to get to ~60 percent of the range; they cannot be competitive with the top, which for them is not affordable. Ontiveros added that they try to incentivize by providing value in the operation, what an employee would bring to the campus, and in offering mobility; also, the benefits are good. Ontiveros emphasized the importance of believing in the mission and being part of a strong culture, with the hope that people coming in will be attracted enough that they will want to stay. He then suggested that if the VA could standardize job titles and then create a similar multilayered approach across the agency, it would offer opportunities and career paths, and enable development of training programs.

WRAP-UP DISCUSSIONS OF MAJOR THEMES FROM THIS WORKSHOP

At the close of both days, the meeting participants addressed possible themes from the workshop. Seeking to consolidate inputs at the end of Day 2, Co-Chair Smith opened discussions on observations prepared by himself, Co-Chair Drury, and Cameron Oskvig, director, Board on Infrastructure and the Constructed Environment. A consolidated list of themes is below. Although there was no universal consensus on themes among participants, especially on specific categorizations and details, the overall message of these themes—much model-related work remains—was fully supported.

Theme 1: Both the VA and private-sector hospitals have dedicated and committed engineering workforces. Presentations and discussions offered ample evidence that both the VA and private-sector engineering staffs are dedicated and committed professionals who believe that they are an essential part of the total facility workforce aimed at guaranteeing quality health care. This is not just a throwaway observation. If there were a division between engineering and the delivery of health care, that would be a concern; the workshop attendees did not see such a division.

Theme 2: A standardized model could have great value for VHA business and planning, but such a model requires substantial consideration of variables. Many attendees believed that there is a model, with variables, despite the cliché, “If you’ve seen one VA facility, you’ve seen them all.” There probably is a set of basics to run a facility, but one must blend with that set other relevant elements of uniqueness, such as those observed during this workshop and illustrated below.

Theme 3: Facility size and location are important parameters. All facilities need a core engineering staff, irrespective of size. In some ways, the demand of running a small facility is no less challenging, and maybe more so, than running a large facility; work at the smaller facility is perhaps being accomplished on a shoestring. Small, remote facilities endure more challenges recruiting, training, and dealing with nonavailability of workforce; they may also have problems obtaining replacement parts rapidly and

hiring contractors locally. In contrast, larger facilities are often in areas with available engineering skills that can be contracted for when necessary. Although larger facilities likely benefit from economies of scale, it is not clear how to measure that benefit in terms of either funding or FTEs.

Theme 4: Specific functions and on-site construction impact O&M in special ways. Conversations about impacts of research on FTEs were intriguing. Organizations with access to research dollars enjoy those dollars being plowed back into either contract support or FTEs, thereby producing advantage over organizations (usually, the smaller ones) lacking access to research; this effect is not yet accounted for in modeling. Materials presented also showed varying FTE ratios for hospital space versus ambulatory-care space versus administrative-support space, thus indicating some kind of relationship between the function and the intensity of O&M efforts. Also, presentations indicated that a facility with a major construction program must live with associated shutoffs, movement of staff out and back into the construction area, isolation of the construction area, and so on. With construction, O&M staff time gets devoted to the coordination needed to allow work by contractors.

Theme 5: Contracted versus in-house service is worthy of much more attention. Whether it is better to contract or to hire depends on a range of factors, such as the type of skills needed and their availability in the nearby area, the comparative costs of contracting versus hiring, rules that favor hiring certain types of contractors, and so on. Sound arguments on both sides are worth exploring in more detail.

Theme 6: Benchmarking FTEs to characteristics in addition to space, such as shop descriptions and the portfolios of various organizations, deserves additional attention. For shop-by-shop benchmarking, the metric of square footage per FTE is a starting point, but it is necessary to drill down to a more detailed look at descriptions of shops within individual engineering organizations. Beyond FTEs, one has to look at the total budget, at various costs, and so on and then build from there.

Theme 7: Performance measurements are necessary to manage. Returning to “You can’t manage what you can’t measure,” there was good discussion of various metrics for measuring performance, some established by using technology, others by considering the results of process improvements, and more by common-sense use of data. Top-level indicators, such as up-time of essential services and costs saved, were presented. Caution was suggested about using the VA’s FCAs, which consider only age, and VA applications of the DoD BUILDER program are still under development.

Theme 8: Management of work orders, which are often incomplete, could probably be strengthened to support improved evaluation of maintenance outcomes and staffing efficiency. Briefings on this subject were very impressive. Many work orders may be self-initiated by maintenance staff listening to customers and observing details as they move around their area. When these are completed, they can be documented, thus providing valuable data. Repair-driven maintenance work orders constitute about half of total time spent, probably more than is optimal. New accounting systems (such as a replacement for AEMS/MERS) could help improve work-order management.

Theme 9: Several additional factors require further committee consideration.

• Training is a key aspect of facilities management. Discussions were especially illuminating with respect to new technology, times and competencies required for certifications, and ways to motivate upgrading of KSAs and entering paths for advancement.
• At all three facility levels, it appeared difficult and time consuming for VHA engineering managers to attract and bring highly qualified O&M candidates on board and up to speed.
• Needs for 24/7 O&M staffing add complications versus O&M required during normal days.
• The NFPA 99 framework (including risk levels, performance-based maintenance) may need to be considered in modeling to ensure its integration or to stay consistent with its intent.
• Growing cybersecurity concerns, especially with Internet-connected systems, are important.

DISCLAIMER: This Proceedings of a Workshop—In Brief has been prepared by Norman Haller as a factual summary of what occurred at the meeting. The committee’s role was limited to planning the event. The statements made are those of the individual workshop participants and do not necessarily represent the views of all participants, the planning committee, or the National Academies. This Proceedings of a Workshop—In Brief was reviewed in draft form by Robert Anselmi (VA Hospital Engineer, retired), David J. Nash (Dave Nash & Associates International, LLC), and Gary M. Thompson (North Carolina Geodetic Survey), to ensure that it meets institutional standards for quality and objectivity. The review comments and draft manuscript remain confidential to protect the integrity of the process. All images are courtesy of workshop participants.

PLANNING COMMITTEE: Colin G. Drury, University at Buffalo, Chair; Robert Anselmi, VA Hospital Engineer (retired); Wesley L. Harris (NAE), Massachusetts Institute of Technology; Gene Hubbard, RiVidium, Inc.; and Fred S. Switzer III, Clemson University.

STAFF: Cameron Oskvig, Director, Board on Infrastructure and the Constructed Environment; Erik Svedberg, Senior Program Officer, National Materials and Manufacturing Board; Joe Palmer, Senior Project Assistant, National Materials and Manufacturing Board; and Daniel Talmage, Program Officer, Board on Human-Systems Integration.

SPONSORS: This workshop was supported by the Veterans Health Administration.

Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2019. Facilities Staffing Requirements for the Veterans Health Administration—Operations and Maintenance of the Physical Plant and Equipment: Proceedings of a Workshop—in Brief. Washington, DC: The National Academies Press. doi: https://doi.org/10.17226/25452.

Division on Engineering and Physical Sciences Division of Behavioral and Social Sciences and Education

Copyright 2019 by the National Academy of Sciences. All rights reserved.