top of page
Search
Writer's pictureJennifer

Haphazard Analysis: A challenge to my peers and colleagues to do better


A Preface: I am passionate about this topic, and the blog length reflects that. I hope you will keep reading!


I fear that doing a hazard analysis has become a rote exercise-- a chore to satisfy auditors, and the occasional regulator. In reality, it is the backbone of your food safety plan and should guide your efforts (e.g., resources) to minimize the greatest risks, and therefore protect public health. I will never forget Warren Stone, while a colleague at GMA, emphasizing to PCQI students that your food safety preventive controls are based on the hazard analysis.


With the FSPCA PCQI for Human Foods Version 2.0 training coming out soon, and a few outbreaks calling into question how food safety is being managed, I feel compelled to share my grave concerns for our industry and profession. Food safety professionals need to do a better job on their hazard analyses. Colleges, universities, and vocational programs need to do a better job training future food industry professionals on the fundamentals, starting with the distinction between hazards and risks, imparting a thorough understanding of each hazard.


Here are a few observations of what I am seeing in industry today:


Confusion over “preventive controls”

o   At first, I thought it was a 1-off. But I’ve come to realize there is a widely held belief that “preventive controls” is the place to list all the Good Manufacturing Practices, prerequisite programs, and anything else one can possibly think of that could have a food safety consequence. No, No, No! When asked “Does this step require a Preventive Control?” and you check “Yes”, you are signing yourself up for the same level of rigor and documentation as if it was a Critical Control Point (CCP). Your preventive controls are the things that matter MOST. They are specific to that hazard, and that process step. A failure in the preventive control means that the food is at imminent and immediate risk. Hairnets are not a preventive control. Be very deliberate as you identify preventive controls- you are going to need to put resources toward them, because food safety hinges on their successful implementation.


Underemphasis on “justify your decision”

o   This is where it’s at! This is the most important box on the table, in my opinion. Too often I see “copy and paste” answers in this box. For example, to justify why a microbiological hazard needs a preventive control the box will say “pathogens can cause illness and death” (and it says that for every “b-biological” box in the table). Ok, but what is the likelihood of a specific pathogen being introduced at this step? Fellow food safety professionals, please take the time to THINK about this and to DOCUMENT this. Let go of the time spent selecting arbitrary likelihood/severity matrixes and spend your time here instead!


Overemphasis on severity and likelihood scores

o   I’m going to go against the grain and encourage deleting the severity and likelihood columns from the hazard analysis. First, there’s little a company can do to change severity—that is generally inherent with the nature of the hazard (we can argue over the relationship between dose and likeliness of acute- or long term- illness, or the size of a foreign material and the associated severity of illness or injury. But I never see companies that attempt to defend their severity scores). Likelihood needs more focus, and this should be extremely well described in the “justify your decision box.”. The likelihood of occurrence is exactly what you can and must control. And likelihood can often be quantified—you should have data to back up your claims of likelihood. Picking a number 1-5, or selecting a high/medium/low is wholly insufficient if the justification of that selection is not robust. I also absolutely hate the likelihood/severity matrix. It is often arbitrary and forces you into a box- literally. It underutilizes the intelligence of well-trained food safety professionals who have the capability of defending their decisions.


CCP designations

o   In light of the fact that preventive controls are, according to regulation, those specific, risk-based programs that are necessary to significantly reduce or eliminate a hazard (e.g., they affect probability/ likelihood so that risk is reduced), why are we still calling out CCPs? I fear I know the answer and it’s unsettling (I’ll save this for another blog). Your CCPs have already been identified- they are in the Preventive Controls column.


Copy and paste

o   The only time I find copy and paste acceptable is when considering the potential for post process contamination with environmental pathogens before a product is packaged. Even then, there may be certain pieces of equipment or certain exposures where the likelihood of harborage (and therefore contamination) is greater.


·Limited use of data or references

o   This is the biggest missed opportunity I see. When evaluating whether pesticides are likely to occur, look at data! Look at USDA and FDA reports (specific for the production region and the crop). When you have access to a data set (e.g., antibiotics in milk, mycotoxin testing, rates of consumer complaints of metal or hard plastic, etc.) use those data- whether internal or publicly available- to inform your assessment of likelihood of occurrence (and cite these data in the “justify your decision” box).


Here is my challenge: if you have anything to do with a hazard analysis, look at it. I suggest you set aside at least 2 hours, ideally a half day. I know how difficult it is to do this with the constant meetings. I would argue that there is no meeting that is more important than having a correct hazard analysis. Go row by row, column by column. Ask yourself these questions:

·       How did the hazard analysis come to be? Did you write it or inherit it? Was it updated because an auditor or regulator had a specific request? If so, was it justified?

·       Are you specific in describing hazards?  “Micro” is nondescript. Bacteria are different from viruses and parasites. Spores are different from vegetative. Growth is different from persistence. Managing Salmonella coming in on produce due to the growing environment is totally different from managing environmental  Salmonella in a dry milk operation.

·       Are you adequately assessing ingredients/ raw materials? When looking down the hazard analysis table, the rows related to ingredients are often the most poorly done. If you’re relying on the supplier to control a hazard (because your process won’t) many supplier approval programs are ill equipped to address this. Consider your supplier approval program. Does it really give you the assurance that things are being carefully managed? How is this verified? Can you trust that verification? When was the last time we saw an outbreak associated with a facility that had passed an audit?

·       Do you have too many preventive controls? If the product will undergo a kill step you don’t need a preventive control to address vegetative pathogens prior to the kill step. Again, the preventive controls should be reserved for those activities that are critical to the safety of the product. This does not mean that programs that support food safety should be abandoned—Good Manufacturing Practices and Prerequisite Programs are important—they are indeed prerequisites—but rarely are they also preventive controls.

·       Are your food safety programs (and the resources to implement them) aligned with your hazard analysis? Here’s an exercise I encourage you and your food safety team to try:

o   Look at all of the rows of the hazard analysis, and especially those for which the box is checked that the hazard requires a preventive control.

o   Write each hazard/step that requires a preventive control on an index card. You may wind up with 3 index cards or 30.

o   Now, place them in order of your concern. If you got a call from a regulator, customer, or team member saying “we have a problem”, where would your mind go before you heard the next line? This probably reflects the issue with the greatest likelihood/ severity combination—the thing that keeps you up at night. Then move on down the list of what you’re most concerned about.

o   Now, compare your order with those of your colleagues. If they differ, discuss why. The information shared may warrant a reconsideration of the “justification” box.

o   Once there is consensus on the order of the items, talk through how the risk is being managed. Are the efforts aligned with the level of risk? Have the designated team member(s) been notified and trained to complete each step properly, including documentation and corrective actions?


To echo Warren’s words, as one wonders what food safety programs to implement, upgrade, or even scale back, ask yourself “What does your hazard analysis say?” If the hazard analysis is wrong, the food safety system is not only indefensible, but may be putting lives at risk.

382 views0 comments

Recent Posts

See All

Comments


bottom of page