Regulatory Open Forum

 View Only
Expand all | Collapse all

ALAP vs. AFAP

  • 1.  ALAP vs. AFAP

    This message was posted by a user wishing to remain anonymous
    Posted 26-Jan-2021 16:23
    This message was posted by a user wishing to remain anonymous

    As we are approaching our MDR audit, our Risk Analysis procedure and documents need to be updated to replace ALAP with AFAP.  

    It seems to be a lot of differences between the interpretation of this change.  Is it fair to say that our organization needs to "demonstrate that risks are reduced AFAP, by showing objective evidence that all possible risk control options were considered and implemented, without taking cost into consideration"?

    If this is true, there is no difference in our organization, from ALAP to AFAP as we never take cost into consideration.

    Please advise.


  • 2.  RE: ALAP vs. AFAP

    Posted 26-Jan-2021 18:48

    There is a lot to unpack here, so I'll break it into the constituent parts.

     

    You say that because of an EU-MDR audit you need to replace ALAP with AFAP. This is the wrong approach.

     

    Under the EU-MDD, many companies implemented EN ISO 14971:2012 which included a (commonly misunderstood) content deviation related to your question.

     

    Under the EU-MDR you need to implement the current version, which is EN ISO 14971:2019. This version does not have any content deviations.

     

    When you look at EU-MDR Annex I, for risk management cite the current version, EN ISO 14971:2019, and the specific clauses. For example, in Annex I(3)(a) "Establish and document a risk management plan" for each device, demonstrate conformity by citing EN ISO 14971:2019 clause 4.4. Of course, ensure that each plan meets the clause 4.4 requirements. They are not the same as EN ISO 14971:2012.

     

    The ALAP v AFAP issue caused a lot of consternation because people didn't understand the difference between acceptable risk and the methods to make the risk acceptable.

     

    ALARP is a set of tools, not an acceptability region. The ALARP tool kit has two major compartments: technical tools and economic tools. The content deviation says you cannot use the economic tools. The paradigm case is the Ford Pinto where the company applied economic tools to determine the cost of the solution to fix a known safety problem outweighed the payouts of potential law suits from occupant death and physical property damage.



    ------------------------------
    Dan O'Leary CQA, CQE
    Swanzey NH
    United States
    ------------------------------



  • 3.  RE: ALAP vs. AFAP

    Posted 26-Jan-2021 20:28
    Edited by Kevin Randall 26-Jan-2021 20:36

    First, it is important to know what an organization's "ALAP" acronym actually stands for.  If the existing Risk Management Procedure defines ALAP as "As Low as Possible", then it may be that the organization's definition of ALAP could already be aligned with the EU MDR's requirement for As Far As Possible (AFAP).  But further detailed assessment of the existing procedure would be needed to flush this out.

    On the other hand, if a Risk Management procedure is based on reducing/controlling risks to be As Low As Practicable (a.k.a., ALARP - As Low As Reasonably Practicable), then, in preparation for an EU MDR audit, it is most definitely the right approach, and certainly required, to revise the procedure so as to update and replace ALAP/ALARP to a proper interpretation of AFAP.  This is because both the outgoing MDD and incoming EU MDR demand that risks be reduced to AFAP.  Indeed, AFAP is not a new concept to the European medical device regulatory arena.

    Ultimately, it is necessary to be sure that the procedure aligns its definition and application of ALAP/AFAP with what will be expected by the Notified Body.  This should start with realizing that the EU MDR's approach to AFAP builds upon, but doesn't clearly conflict with, that which was already established in the MDD.  Specifically, the EU MDR adds that the reduction of risks AFAP means reduction of risks AFAP without adversely affecting the benefit-risk ratio.  So, what does that mean, and how do we characterize it for the purposes of the Risk Management procedure and process?  I introduce this below.

    First, remember that the notion of achieving AFAP without regard for economic considerations was born from Annex Z Content Deviation #3 of EN ISO 14971:2012 (now withdrawn and superseded by the latest state of the art, EN ISO 14971:2019) which includes no such Content Deviation.  Indeed, even before EN ISO 14971:2012 was withdrawn, the Notified Bodies recognized and voiced such recognition (based on longstanding MDD precedent) of the impracticality and inappropriateness of pursuing AFAP with total disregard for economic considerations.  So, the Notified Bodies in fact continued to permit a reasonable amount of economic consideration to be made when pursuing AFAP.  Accordingly, the EU MDR Risk Management procedure's attendance to AFAP should not be singularly focused on an aversion to, or disregard for, the associated economic considerations.  Such an approach would probably be viewed by the Notified Bodies as impossible, and perhaps even disingenuous.

    Instead, the Risk Management procedure needs to assure that the organization and its risk management mechanism properly and precisely define, in practical (i.e., working, implementable) terms, the concept of the AFAP threshold and how the organization goes about reducing risks AFAP.  The procedure needs to specify what particular thresholds and criteria will be used for determining whether a particular risk has or has not been reduced to the AFAP threshold.  The Notified Bodies, based on longstanding precedent, have a pretty clear understanding of this concept as elaborated on in their October 2014 Consensus Paper for the Interpretation and Application of Annexes Z in EN ISO 14971:2012.  As noted above, the EN 2012 version of the ISO 14971 standard has been superseded, yet the 2014 Consensus Paper still in my experience contains an accurate Notified Body interpretation of the AFAP concept to be used as the basis for the corresponding elements of the Risk Management procedure.

    A quick final note:  Risk management is not risk analysis (though risk management certainly includes risk analysis).  Accordingly, in order to comply with EN ISO 14971 and the EU MDR, it is necessary to have a Risk "Management" procedure that properly defines and integrates risk "analysis".



    ------------------------------
    Kevin Randall, ASQ CQA, RAC (Europe, U.S., Canada)
    Principal Consultant
    ComplianceAcuity, Inc.
    Ridgway, CO
    United States
    www.complianceacuity.com
    © Copyright 2020 by ComplianceAcuity, Inc. All rights reserved.
    ------------------------------



  • 4.  RE: ALAP vs. AFAP

    Posted 26-Jan-2021 20:34
    Made a few clerical corrections a couple minutes after posting; be sure to read my latest narrative.

    ------------------------------
    Kevin Randall, ASQ CQA, RAC (Europe, U.S., Canada)
    Principal Consultant
    ComplianceAcuity, Inc.
    Ridgway, CO
    United States
    www.complianceacuity.com
    © Copyright 2020 by ComplianceAcuity, Inc. All rights reserved.
    ------------------------------



  • 5.  RE: ALAP vs. AFAP

    Posted 27-Jan-2021 03:47
    Hello,

    Remove references in your risk management process to ALAP and AFAP.  In your risk management process utilise ISO 14971/ISO-TR 24971 to define acceptable risk and how you determine acceptable risk - this can vary depending on product type so this should be reflected in risk management planning.  It has nothing to do with cost, or as low as possible, or as far as possible - it is about what risk are acceptable and which are not.  There are way, way too many people confused on ALARP, ALAP, AFAP, etc., so make it simple to only speak about how you accept risk and control the risk associated with the product.

    Then the second portion of this is make sure to apply risk control to each hazard, just not those which are above your defined acceptable risk level.  (At least this should be the approach in the European Union, other regions if you already have acceptable risk may not require further risk control.)  By defining and showing whatever risk control can, should, and would be applied in some form of essence this shows risk being reduced as far as possible.

    ------------------------------
    Richard Vincins RAC
    Vice President Global Regulatory Affairs
    ------------------------------



  • 6.  RE: ALAP vs. AFAP

    Posted 27-Jan-2021 07:57
    Edited by Mark Swanson 27-Jan-2021 07:58
    This discussion continues to bring out passionate comments.

    There is no black line between what is acceptable or unacceptable. Simply apply the current state of the art (what you can legitimately do) to reduce ALL risks as low as you can. Once you do that, the risk is acceptable. You CANNOT stop applying risk reductions. As your capabilities improve (your state of the art improves) or you discover new/changes in risk, you must continue to apply reductions that you can do.

    Risk management is preventive action (actions you take to prevent occurrences of failures) if you do this without stopping, your product is improved and your cost of poor quality is reduced (saving your company money)--this is also required per ISO 13485/QS Regulation.

    The problem that comes in is when manufacturers release their product and then stop applying risk management. The risk management file (as part of the overall design history file) goes into a cabinet and is no longer maintained.

    You must apply post-market surveillance to continue to identify issues and reduce risk (NEVER STOP). What is acceptable today MAY not be acceptable tomorrow. Yes, this means you must have part of your organization dedicated to maintaining legacy product and continuing to work to reduce (and maintain) the acceptable risk. GOOD LUCK!

    ------------------------------
    Regards,
    Mark Swanson, ASQ CBA, CMQ/OE, CQE ASQ, MBA
    Becker MN
    United States
    ------------------------------



  • 7.  RE: ALAP vs. AFAP

    Posted 27-Jan-2021 11:58
      |   view attached
    A whole lot going on here with plenty of misunderstandings over the years. I wrote an article on Risk Acceptability for MeDevice Online a short time ago, which I will attach here.

     As was indicated in previous posts, EN ISO 14971:2012 was withdrawn by CEN and CENELEC and replaced by what BSI has indicated is the "state of the art" medical device risk management standard.  Also you cannot use any Harmonized standard currently, as there are no standards Harmonized to the MDR / IVDR.  Also a prefix of EN does NOT mean a standard is Harmonized.  Standards are no Harmonized until the EU issues a Standardization Request to the European standards bodies CEN and CENELEC and they create a standard that is published in the Official Journal, indicating it has been accepted as Harmonized.  To date one such request was submitted to CEN and CENELEC and rejected by them.  Another was in the process of being created for submission last month.  Currently ISO 14971:2019 and EN ISO 14971:2019 are identical with no differences.  When Harmonized an Amended version of EN ISO 14971:2019 will be released with the Correspondence Tables for MDR and IVDR (an maybe even the Directives, though they might be useless at that point).

    Your current choice is to either create your own system or to use EN ISO 14971:2019 to comply with risk management requirements in the MDR / IVDR.  You may then comply with the standard, and respond to any requirements of the MDR / IVDR that are not covered in EN ISO 14971:2019.  The ISO FDIS 14971:2019 contained proposed Correspondence Tables for the MDR (ZD) and IVDR (ZE) that show how such a table might be constructed.  Unfortunately without Standardization Requests for any standards being accepted, these tables are only samples and do not contain valid data.  The new Z Annexes will no longer have the troublesome Content Deviations that were alluded to early.  They were fraught with errors which were addressed in the new regulations and also addressed in the 2019 version of 14971.

    Now as for the ALAP / AFAP debacle, those are possible choice for a risk management POLICY as required in ISO 14971:2019  4.2, and are NOT regions on a risk chart.  EN ISO 24971:2020 4.2 and Annex C go to great lengths in discussing the differences between these commonly confused requirements.  The referenced article discusses this.  The Policy sets what the company will use and consists of: Purpose; Scope; Factors and considerations for determine acceptable risk; Approaches to risk control; and Requirements for approval and review.  The Risk Acceptability Criteria should be based on (and here is where MDR / IVDR comes in): Applicable regulatory requirements; Relevant internal standards; Generally acknowledged state of the art; and Validated stakeholder concerns. One of my concerns was always how you prove you have reached As Far As Possible, there,  is always one more risk control that can be applied in my estimation, now the trick is to make sure you did not impact benefit-risk, that is your out, making the product unavailable due to cost is one impact that is certainly negative.

    As Kevin indicated, once you have established a company Policy for Risk Acceptability, then for a particular product, you establish Risk Acceptability Criteria for that product in the Risk Management Plan for that product.  Those criteria help you to implement your policy (ALAP, AFAP: without impacting your benefit-risk ratio; or whatever it is).

    One of my other concerns is the undefined terminology used in MDR / IVDR.  Benefit-risk ratio seems to indicate a numerical value, which cannot be created with all the disparate elements of benefit and risk, and then trying to compare them.  EN ISO 24971:2020 does provide some examples on documenting benefit-risk, but they are all textual and not numeric, so there is no "ratio".

    Good luck on your quest, but in the end your Notified Body will decide if you have reached your goal, or not.


    ------------------------------
    Edwin Bills MEd, CQA, RAC, BSc, CQE, ASQ
    Principal Consultant
    Overland Park KS
    United States
    elb@edwinbillsconsultant.com
    ------------------------------



  • 8.  RE: ALAP vs. AFAP

    Posted 27-Jan-2021 13:31
    Edited by Kevin Randall 27-Jan-2021 13:44

    When approaching risk management to comply with the corresponding requirements of the EU MDR, I feel it is imperative to remember that the EU MDR, in more than twenty different instances, legislatively demands that risk be reduced "as far as possible" (AFAP).  Another telling factor relates to the aforementioned EU MDR definition of AFAP, which means reduction of risks AFAP without adversely affecting the benefit-risk ratio.  Specifically, this identical paradigm is stated by EN ISO 14971:2019 under clause 4.2 therein, reminding us that the manufacturer has the liberty to define its particular approach to risk control, and that such an approach could (but for the legislative context of the EU MDR, must) be based on reducing risk "as far as possible without adversely affecting the benefit-ratio".  Accordingly, when applying EN ISO 14971:2019's normative provisions in the context of the EU MDR's legislative mandate for reduction of risk AFAP, I suggest that the abandonment of the AFAP concept in an EU MDR risk management system will likely result in objection from the Notified Body, probably in the form of a major nonconformity.

    Offline I received an objection from a peer regarding the notion of establishing an AFAP "threshold".  The concern was lodged in response to my preceding post, where I stated that the EU MDR Risk Management procedure needs to assure that the organization and its risk management mechanism properly and precisely define, in practical (i.e., working, implementable) terms, the concept of the AFAP threshold, and how the organization goes about reducing risks AFAP.  I also stated that the procedure needs to specify what particular thresholds and criteria will be used for determining whether risk has or has not been reduced to the AFAP threshold.  The objection asserts a concern that there is no such threshold in ISO 14971, and asserts that we should abandon such a notion.  The objection also asserts that we should instead focus on acceptability without any threshold, and that there is no black line between acceptable and not acceptable.  My approach to these concerns is described below.

    Specifically, if we reject the notion of an AFAP "threshold" (the Notified Bodies call it an AFAP "end-point") and thus try to tackle the AFAP beast without first establishing implementable, executable boundaries on the meaning of AFAP, then it would seem to intrinsically demand and prevent us from ever concluding that the device's current risk profile is acceptable at any moment in time.  Indeed, a key factor motivating the withdrawal of the 2012 EN ISO 14971 Annex Z Content Deviations on AFAP was the Notified Bodies' recognition that, without a "clear, easy to understand, and unambiguous" AFAP end-point, there would be "practical problems such as where to stop in reducing risk".  I'm suggesting that this problem would only be exacerbated (and we would likely ruffle our Notified Bodies feathers) if we were to reject the notion of establishing an AFAP "threshold" (i.e., end-point) for EU MDR risk management.

    As mentioned in my prior post, the Notified Bodies seem to have a clear understanding of the AFAP end-point, and it is explained in the aforesaid Notified Body Consensus Paper.  Moreover, I find that there is a clear correlation between that approach and how risk acceptability is addressed in ISO 24971:2020 (the guideline for ISO 14971), thus providing a nice bridge between the EU MDR's legislative parameters and ISO 14971's normative parameters.  This is a good opportunity to mention that Edwin Bills' article (https://www.meddeviceonline.com/doc/iso-tr-bringing-clarity-to-risk-acceptability-in-iso-0001) is a great read.

    In summary, the Notified Bodies' approach to establishing the AFAP end-point is what needs to be woven into an EU MDR Risk Management procedure/process.  Via this approach, I have had positive feedback and results during Notified Body audit of my Risk Management system. But if the AFAP approach and establishment of a corresponding end-point are abandoned regarding EU MDR conformity, then my opinion is that a major nonconformity from the Notified Body would follow soon thereafter.



    ------------------------------
    Kevin Randall, ASQ CQA, RAC (Europe, U.S., Canada)
    Principal Consultant
    ComplianceAcuity, Inc.
    Ridgway, CO
    United States
    www.complianceacuity.com
    © Copyright 2020 by ComplianceAcuity, Inc. All rights reserved.
    ------------------------------



  • 9.  RE: ALAP vs. AFAP

    Posted 27-Jan-2021 15:22
    Kevin,

    You captured my concerns with better words.  I have been concerned with trying to find that objective evidence to prove reaching AFAP.  And then trying to come up with a "benefit-risk ratio".  No way to do that I know of, as there are too many disparate factors in benefit to come up with numeric values for a ratio, to say nothing of overall residual risk.

    Anyway, I want to commend you on your excellent post, you have captured one of the most problematic areas of the MDR.

    ------------------------------
    Edwin Bills MEd, CQA, RAC, BSc, CQE, ASQ
    Principal Consultant
    Overland Park KS
    United States
    elb@edwinbillsconsultant.com
    ------------------------------



  • 10.  RE: ALAP vs. AFAP

    Posted 28-Jan-2021 08:10
    I too commend Kevin for his posting. This captures a couple of great points that continue in Ed's post that can be difficult.

    1) You can't put a number to risk (it cannot be calculated). Despite the efforts of many companies to make this completely objective using reliability numbers, complaint numbers, opportunity numbers (sales or some derivative thereof) that really doesn't work. You must have the competent experts to judge the acceptability of the residual risk. I find it is relevant that the section on competence of personnel is right behind the section on top management establishing the policy on risk acceptability. This gets even more attention as this version of 14971 has a new definition for "benefit". Thus (as Ed points out) this need to have a benefit/risk ratio has some meaning, however when you are talking to a group of engineers and using a term like 'ratio', it makes sense that the audience wants to make that into a number (again as Ed outlined...how do you come up with numbers for each of these?? Thus the difficulty in explaining all this to engineer types.

    ------------------------------
    Regards,
    Mark Swanson, ASQ CBA, CMQ/OE, CQE ASQ, MBA
    Becker MN
    United States
    ------------------------------



  • 11.  RE: ALAP vs. AFAP

    Posted 29-Jan-2021 18:11

    All the theory notwithstanding, for a device manufacturer the question is implementation. I contend it is straight forward.

    My scope is the MDR Annex I using EN ISO 14971:2019 and the concept of risk reduction. There are other issues, such as disclosure of residual risk, excluded from this scope. I also use a common approach to the risk matrix, but it not required.

    Set a policy for risk acceptability, such as reducing risk as low as reasonably practicable (ALARP), reducing risk as low as reasonably achievable (ALARA), or reducing risk as far as possible without adversely affecting the benefit-risk ratio (AFAP-BR).

    Define the risk matrix. Tradition has five levels of severity and five levels of frequency of occurrence resulting in twenty-five cells. Analyze each hazard (hazard, sequence of events, hazardous situation, the harm, the harm's severity, the harm's severity's frequency of occurrence, risk estimate). The (initial) risk estimate is the specific cell in the risk matrix.

    The manufacture's risk acceptability policy defines the boundaries between the regions.

    Risk evaluation determines the need for risk reduction. Tradition assigns each risk matrix cell an attribute that also implements risk evaluation. ISO 14971:2007 suggests three risk regions.
    Unacceptable
    Acceptable with risk minimization
    Acceptable – Negligible

    Risk reduction employs a combination of the hierarchy and the tools. In simple terms, the hierarchy is: design out the harm, protect against it, or warn about it. Simply, the tools are technical means and economic means.

    After implementing the risk control measures, evaluate the residual risk using the risk matrix. Risk reduction brings the residual risk to the best cell by implementing risk control measures based on the hierarchy. The best cell is the lowest severity and the lowest frequency of occurrence. In a properly constructed risk matrix, this is as close as one can get to the lower left cell.

    The EU-MDR has requirements that override some of EN ISO 14971:2019. For example, Annex I(2) requires reducing risk as far as possible without adversely affecting the benefit-risk ratio (AFAP-BR) for those hazards identified in Annex I. Annex I(2) says, "The requirement in this Annex to reduce risks as far as possible means the reduction of risks as far as possible without adversely affecting the benefit-risk ratio". In practice, apply the criterion to all hazards regardless of the source.

    Notice that the risk reduction criterion is only the benefit-risk ratio, not technical means, economic considerations, etc.

    Unfortunately, the EU-MDR does not define "benefit-risk ratio". However, it is easy to infer the meaning. At the top level of the device, there are benefits to using the device and risks to using the device. Think of a pan balance (a type of scale) to describe the ratio; the benefits must outweigh the risks. (The benefit pan is lower than the risk pan.) (See also Art. 2(24) for the definition of benefit-risk determination.)

    In practice, the overall benefits. define the overall risks, and form the ratio. The ratio should start in the patient's favor.

    Then, as you implement the risk reduction measures, check for two things. Is that residual risk in the best possible cell? Do the risk control measures have an adverse effect on the benefit-risk ratio?



    ------------------------------
    Dan O'Leary CQA, CQE
    Swanzey NH
    United States
    ------------------------------