The simple answer to the question posed in the title of this article is, "it depends." Many years ago, my children teased that the domain name for my consulting practice should be "itdepends.com" because that is how I answer every question. However, in this case, there is no simple answer. The questions about the safety of controlled environment agriculture (CEA) are complicated. This article outlines why CEA safety depends on understanding and properly addressing the challenges of combining agricultural and ready-to-eat (RTE) into a single facility. It also addresses why CEA is not inherently safe. Ultimately, this article will show how the risk profile of CEA must be compared to that of other RTE produce and that there is no universal answer to the question of CEA safety.

CEA is the flavor of the moment for investors and technocrats who believe they have a better way to produce food. It is asserted to be safer, more sustainable, more controlled, free of pesticides, local, fresher, have a lower carbon footprint, and more. Many of these claims have some basis, but they are broad generalizations that are not uniformly true. This article does not address all of these claims. Instead, it focuses on the food safety aspects of CEA, which is still a complex matter. To aid the reader, this article is divided into discussions of the following questions:

• What is and is not CEA?
• Is CEA ripe for an outbreak similar to that seen in 2006?
• What are the key challenges of melding farming and food processing?
• What is the path to safety for CEA?

What Is and Is Not CEA?

CEA is a strange concept. All agriculture requires a controlled environment. Plants are not motile and cannot move away from problems. Plants have many environmental requirements to grow. Traditional farmers, those growing outdoors, must protect their crops from many threats. They are largely successful, or there would be no harvest. However, the term CEA has come to refer to operations that are much more contained, particularly those that do not use soil and are in a building. Some CEA operations use peat; others use rock wool. Most CEA operations use water-delivered nutrients. Some traditional farmers use water to convey nutrients, too. Some CEA operations provide artificial light, and some provide enhanced carbon dioxide. Traditional agriculture must cope with weather, but this leads to various climates being selected for specific crops.

Weather elimination and enclosed buildings are probably the most distinctive feature of CEA efforts. The intent is to yield year-round crops. Enclosure of the operation leads to the belief among many consumers that CEA prevents interaction with pests and human pathogens to provide a "safe" (as opposed to "safer") product.

The vision consumers have of CEA is often more like that of a biodome than a simple greenhouse. A biodome is a closed habitat that is built to represent a particular environment and the community of organisms that live there. Ideally, a biodome will require no inputs except light or energy for artificial light and have no outputs. This level of containment is impossible for a CEA operation. CEA operations generally have soil substitutes and nutrients as inputs in addition to light or energy to power artificial light sources. The product is clearly an output, but there are usually byproducts such as the used substitute soil and semi-consumable supplies such as gutters or floats, which are used to support the plants. Even the most automated CEA operations require some people, which limits the achievable degree of closure (Figure 1). Human intervention will be part of both farming and processing operations in CEA facilities. Foot traffic has been shown to track Salmonella into a CEA greenhouse.

The level of control that is achieved in agriculture generally, and in CEA in particular, is quite variable. The range of control starts with open fields placed for some measure of control and extends to hoop houses, to greenhouses with windows that open, and then onto more closed environments, as are often found in vertical farms where all of the lighting is artificial. Ultimately, the marketplace will establish the value of the various levels of isolation or control based on the perceived value of the product. The safety of products will play into this valuation. A partial list of different types of CEA categories includes the following:

• Greenhouses
• Hydroponics
• Aeroponics
• Vertical farms
• Aquaculture
• Aquaponics.

Given that CEA operations are inherently partially open, it is important to consider what is necessary to mitigate potential hazards to maximize safety. This maximization of safety will require risk assessment on an operation-by-operation basis. The categories of vectors to be controlled are surprisingly similar for CEA and traditional farming when viewed in the light of the claims made about CEA. All farming operations need to consider water, wildlife, environmental factors, workers, and others as potential vectors. This list can be parsed in many ways, but a high-level examination of some categories will illustrate the similarities.

Traditional farming must properly place fields to partially mitigate air as a vector. If there is a known hazard, it is often better to relocate to a different field. Air exchange in CEA operations can make air a vector. It would be possible to use clean room technology to eliminate air as a vector, but the costs could render CEA impractical. Potentially, there is a middle ground. Risk assessment analysis is required to identify the hazards and risks associated with air as a vector for both CEA operations and traditional farming.

Water is another potential vector. Traditional farming is moving toward better control of this input with greater use of deep wells and water treatment. Agricultural water has been linked to multiple outbreaks including those associated with CEA. In the process environment, water chemistry is critical for safe operation. The control of process water chemistry has evolved with many automated dosing and control systems allowing high levels of performance and verification of performance. CEA will need water inputs. The quality of the water used to carry nutrients to the plants is of concern. The recirculation of this water presents its own challenges with the potential to allow pathogen growth. Incidental contact with edible portions of the crops is of concern. At present, there is much discussion regarding whether and how to wash CEA product. Ultimately, these concerns must be addressed as part of a complete risk assessment profile. This assessment will need to be carried out for each operation to the extent that operations are different. Again, water must be considered as a vector for both CEA and traditional farming operations.

Amendments and seed are other vectors that both traditional farming and CEA must address in their risk assessments. Seeds are difficult to sanitize. Sterilization of seeds to prevent inoculating doses of undesirable pathogens will damage germination rates or, in some cases, prevent germination. The microorganisms on seeds are largely a fact of life for all agriculture. The other types of additives are largely specific to the type of agriculture. In CEA, soil substitutes are an important potential source that must be examined. Traditional farming is increasingly focused on soil amendment quality. Compost, chicken pellets, and fish emulsions have and are receiving scrutiny. Clearly, all agriculture operations must assess the risks associated with these materials and take appropriate precautions.

The impact of animals is probably the area of greatest difference between traditional farming and CEA. A building will provide better control than fences and isolation. Traditional farming continues to work on this area with various tools including setbacks, relocation, rejected product, and greater vigilance. These methods will never be perfect. Birds are the biggest remaining challenge. CEA has an advantage in the closed building; however, the impact of large animals can be mediated by air, where CEA is still vulnerable. CEA operations must consider their land neighbors.

CEA operations are trying to appear more "wholesome" than traditional organic, presenting themselves as chemical and pesticide free. Insects are challenges for all agricultural operations, including CEA. I have witnessed fly infestations in CEA operations. Insects can be carriers of plant pests and human pathogens and can do direct damage. It appears that CEA operations must implement some manner of insect control, as CEA is not closed enough to exclude the mitigation of hazards associated with insects.

Although not a safety concern, the nutrient formulations will, in general, contain chemical components that are not on the U.S. Department of Agriculture's National Organic Program (NOP) list. At some point, consumers will become aware of the use of these chemicals and may feel deceived. Eventually, the marketplace will render judgement.

Bottom line, the vectors for CEA are largely the same as for traditional agriculture, with some differences. These challenges are not insurmountable, but planning and execution are required. CEA operations must build on the foundations of traditional agriculture. As they say, there is no "free lunch." Safety can be achieved only with attention to the details.

Is CEA Ripe for an Outbreak?

One would like to think that the U.S. Food and Drug Administration (FDA) and the Centers for Disease Control and Prevention (CDC) would take a more targeted approach to a produce outbreak and not shut down an entire sector of the ready-to-eat (RTE) marketplace, as was done in 2006. However, in 2018, an outbreak event was attributed to romaine lettuce, and the market sector was essentially shut down. One can only speculate about the potential for CEA to be excluded from a general shutdown of a product category when or if a potent outbreak occurs. One can also speculate about FDA shutting down the entire CEA category if outbreaks continue. FDA does seem to have moved away from this type of response; however, FDA has also pointedly asserted, in its assessments of the CEA-related outbreaks that have occurred, that CEA is not inherently safe.

Returning to our question, in 2006, the RTE spinach industry was shuttered because of contamination with pathogenic E. coli. Hundreds of people got sick while the industry provided hundreds of millions of servings. The increased consumption of raw spinach clearly played a role. The whole story is much too long to rehash here; however, three aspects stand out as relevant enabling factors for this outbreak that may apply to CEA. First, enough illness occurred to provide enough product type-specific traceback that the regulators felt compelled to act. Second, prevailing consumer attitudes were that washed spinach was safe as long as the dirt was removed. Third, the product volume was large. I am sure some would say this is a naïve analysis and point out that many more technical factors were in play. However, I would ask that they bear with me for a moment while I expound on these ideas.

It is probable that commodity spinach has always caused illness at some low level when consumed raw. However, such illnesses disappeared into the background as "unexplained foodborne illnesses." The rise of branded RTE products made the product category a target and prompted the linking of multiple cases together to identify at least a partial source or cause of the illnesses. The ability to trace back has continued to improve in the intervening years. CEA product seeks to be distinctive and, therefore, will be readily identified in traceback efforts. Outbreaks and recalls associated with particular CEA products have already occurred.

As part of its outbreak investigation, FDA concluded that many contamination risk factors for CEA are similar to those found in traditional, open-field agriculture. FDA highlighted the following requirements and recommendations for CEA growers:

• Assess and understand potential sources and routes of contamination within and around operations
• Implement effective sanitation procedures and sampling plans
• Apply the Food Safety Modernization Act (FSMA) Produce Safety Rule and Good Agricultural Practices (GAPs)
• Maintain proper temperature
• Ensure that all water used in the operation is of adequate water quality
• Perform root cause analysis (RCA) and implement appropriate preventive measures.

Unfortunately, my experience with CEA practitioners is not encouraging with regard to risk mitigation efforts. In my view, CEA growers assume that because they do not have dirt and are enclosed, they are safe. This attitude seems similar to the prevailing attitudes in 2006. The CEA industry needs to absorb the learnings of more than 30 years of traditional farming related to RTE foods in a short time to avoid problems. This is a broad, negative characterization of CEA. Some CEA players are probably well down this path; time will tell. However, the blanket characterization of CEA as safe is unfounded. There are a number of confirmed risks that need to be mitigated.

In spite of the hype and the rapid growth, CEA remains a small portion of the RTE industry. Given that large outbreaks are primarily a numbers game, CEA is unlikely to be the focus of an industry-stopping outbreak. All bets are off when there is an "oops" moment and someone seriously increases the risk of illness. However, barring undesirable acts of commission or omission, the numbers do not support the idea of CEA shutting down a portion of the RTE industry. It might be instructive to examine the number of reported illnesses per serving sold—this would normalize the comparison. Unless CEA operations can turn the corner on profitability, they are unlikely to achieve enough volume to affect entire sectors of the RTE marketplace.

CEA probably cannot cause a 2006-like event. However, CEA needs to do better as it expands. The marketplace will forgive some early mistakes as CEA gets started; however, it is probable that continuing problems would swamp the influence of marketing claims and stifle the growth of CEA.

Key Challenges of Melding Farming and Food Processing

The biggest challenge for CEA operations in melding farming and food processing is that they still have all the same challenges associated with traditional, more separate operations. As discussed above, many of the same vectors need to be considered in each half of the operation that are found in their traditional counterparts. These challenges are made more difficult by proximity. CEA business models that rely on reduced human labor may come up short. CEA operations do not have the same advantages of scale to spread overhead costs as do their larger cousins in the traditional RTE production system.

Having both farming and processing operations under one roof leads to challenges related to proximity and the need for separation. Processing safety devolves into having a high-care area where clean product is packaged. This requires traffic control. People traffic should be highly restricted. In traditional processing plants, this separation is often achieved by using the wash system to transport product from the receiving side into the high-care area. If no wash system is installed, then some other "one-way" door is needed to act as a pass-through. If intact product from the farming operation is transported into the high-care area, then the potential for contamination increases. Some operations may use a cordoned-off area of the high-care room for the initial processing of the fresh-from-farm product. This separation must be carefully examined as part of the risk assessment.

Having a farming operation onsite does not eliminate the general requirements for the processing side of the operation. Traditional plants have recognized the need to treat their operations as food processing facilities. Food processing plants have developed aggressive environmental monitoring problems where potential problem areas are sampled by swabbing on a regular basis. CEA processing will need to follow a similar path and consider many details:

• Nightly third-shift sanitation
• Self-administered Good Manufacturing Practice (GMP) audits
• Washing, wash chemistry control, and drying of product
• Verification and validation of sanitation practices
• Sanitary design and maintenance of equipment
• Material handling, especially packaging, to protect the high-care area
• Employee-related sanitation concerns.

Most, if not all, traditional RTE operations are cold, wet environments. Listeria monocytogenes is of particular concern. Salmonella is generally of greater concern in drier operations. Given that CEA operations can have both wet and dry areas, it is reasonable to consider the development of tailored environmental monitoring programs that address both hazards.

Control of the wash system in a traditional RTE processing facility is an important focus of the food safety program. Washing value-added product has reduced illnesses. CEA, however, has largely bypassed the washing step. Some CEA operations are grappling with the decision to wash or not wash. A handful of operations are washing as a point of differentiation. Since 2006, the science and technology behind washing RTE product have advanced tremendously. Each CEA operation must consider how the safety of its system compares to the current standards of the RTE industry as a whole to support its claims of safety.

CEA farming presents some unique food safety challenges that traditional operations do not face. The recycling of nutrition solutions as part of irrigation has the potential to contaminate an entire production facility. Given the tendency toward continuous production, there are no discrete lots or clean breaks when a problem occurs.

Food safety needs to be designed into operations. The current designs of CEA operations are tailored for optimum operational performance. These designs will evolve in the same way that traditional process plants evolved. Traditional operations are no longer packing sheds. This evolution has taken time. However, given the achieved safety in traditional operations, CEA operations will need to evolve faster and make fewer mistakes.

The bottom line is that bringing the farm and the processing plant together changes the problems, but does not radically simplify the challenges of safe operation.

What is the Path to Safety for CEA?

The path to safety for CEA looks a lot like the path for traditional operations. In the three discussions above, risk analysis is presented as the panacea for all of the food safety challenges. Unfortunately, the ability to execute a complete risk assessment is still evolving. The standards for risk assessments in the traditional RTE sector are soft. Third-party audits are very prescriptive and often focus on minor details without addressing the fundamentals that must be improved to advance food safety. Given the developmental stage of CEA, it will be difficult to execute a complete risk assessment of CEA operations at present. I expect someone will attempt a risk assessment of a farm in an RTE plant in the near future. It will probably be denounced and attacked; however, it will provide a starting point for a much-needed effort.

If such an assessment is completed, it is probable that many fundamentals in the operation will be found lacking and require costly changes. The ability to perform these risk assessments will improve if the CEA community is willing to embrace change and acknowledge that CEA is not inherently safe. The various approaches to CEA will be sifted and sorted by the marketplace. The best processes, and those that evolve in the safest and most profitable manner, will survive.

With CEA operations clawing their way to profitability, cost-cutting is the name of the game. It is important that food safety be considered a fundamental foundation of the business and not a luxury insurance program. The traditional RTE sector invests heavily in food safety and continues to seek improvement. CEA operations need to invest in food safety to survive. They must gather the necessary technical expertise to address the known risks. Every food company is just one or two outbreaks away from failure. CEA operations face all the difficulties of small farms and small food processing operations, requiring the expertise of capable people wearing many hats.

Without the benefit of a complete risk assessment, it is possible to identify a few large opportunities for the CEA sector to move down the path of safety:

• Either employ washing, or demonstrate that washing does not enhance consumer safety
• Develop and implement ways to prevent pathogen proliferation in recycled nutrient systems
• Implement a lot system with clean breaks in the farming operations
• Treat pathogen events as learning opportunities by performing root cause analysis and developing improved processes to prevent similar events
• Go to school on traditional food safety programs and implement the appropriate programs.

CEA represents an exciting, potential compliment to the existing offerings of the fresh, RTE world, but CEA is far from mature. There will be growing pains. Consumers seem willing to pay premium prices for these products, but growth will be sustained only by delivering against the promised and perceived benefits. Any agriculture is only as safe as the food safety programs associated with it. FDA made the point that CEA is not inherently safe. It will fall to the practitioners to ensure that consumers receive the promised value.