Skip to content
HearingTracker
Log in

Importance of REM

74 replies339 viewsLast reply

Jump
Link to this topic
Reply

In reply to @Hearing_Potential

This topic has come up before and the dust settled with what @Um_bongo is saying. I brought this up with my audiologist and she was barely able to measure with phone in help to the point I'm not sure the output is correct. That's for the pediatric team supposedly. Clearly she and every other audiologist I've seen thinks is accurate or good enough. But what @Hearing_Potential notes makes sense unless the system is already doing each of the steps within the REAR calculation. @Hearing_Potential, have you observed a different REAR result with and without patient specific RECD?

ChatGPT created an example where this may matter clinically, but I'm not able to assess the accuracy.

Suppose the fitting software calculates targets using an average RECD, but the patient's actual RECD differs by 8 at a particular frequency. Depending on the fitting prescription and audiometric transducer, the estimated ear-canal threshold and therefore the target may be wrong by a clinically meaningful amount.

A direct REAR could then be measured perfectly and still be compared against a target that is not optimally individualized.

There is an important qualification: this error is not inevitable. Some REM systems use appropriate transducer-specific corrections, convert thresholds using measured RECD data, or derive targets through other methods. You need to know how the particular system calculates its targets before concluding that measuring RECD will improve them.

It seems there may be incremental value depending upon the workflow in that the targets may be improved. I assume not every REM system works the same and possible some convert audiometric thresholds from to estimated ear-canal using an average RECD such that the target itself may be based on an imperfect estimate of the patient's individual acoustics. Even a perfectly executed direct REAR measurement would not correct for an inaccurate target on some or possibly all REM systems?

This source seems to support @Hearing_Potential:

Taking all of these into consideration, verification systems employ prescriptions (, NAL) that use the RECD when it is appropriate to do so, and automatically substitute age- and coupling-appropriate RECDs when necessary. Audiologists can support the accuracy of this process by providing measured RECDs when possible. Accurate incorporation of the RECD, when measured using the same transducer coupling (earmold or foam tip) as is used in threshold measurement, is expected to produce a more accurate transformation from to SPL thresholds which in turn produces more accurate output targets which are based on the SPL thresholds. This allows for an accurate estimate of the hearing aid output that will be achieved at the individual child’s eardrum. These transforms are automatically performed by pediatric hearing aid fitting software systems that employ either DSL v5 or NAL-NL1 or NL2.

Pg33 at https://www.audiology.org/wp-content/uploads/2021/05/PediatricAmplificationGuidelines.pdf1

There is some favorable research about RECD, but it is not conclusive as to this issue. https://pubs.asha.org/doi/abs/10.1044/2022_AJA-21-00264

In the endless REM debates, it seems that audiologists who do not perform REM should be measuring RECD, which they likely are not doing. Now there's an even tougher challenge to find a provider who will do REM and measure RECDs.

@user490 said:

@Hearing_Potential, have you observed a different REAR result with and without patient specific RECD?

The prescription targets differ when a measured is used, so the final will be different by design if your clinician is matching to target. Every system using fitting formulae from NAL or will incorporate the average RECD if an actual RECD isn't available. If your fitting targets aren't accurate in the first place because of the use of an average RECD, the starting point of a fitting won't be as optimal as it could be.

We've spent a lot of time researching this area because infants are depending on what we do for their speech, language, cognitive and social development and growing ears mean that we're working against moving targets so need to achieve an optimal result with each new earmold/refit because it won't be optimal for a considerable proportion of the time the child is using that particular configuration before the next earmold is needed because the ear has physically grown so much.

Adults are a more stable target but we should take the same level of care in my opinion, because they also deserve the best we can give them.

Are the fitting targets generated by the hearing aid software or also by the software? It may not matter if neither know the measured .

Anyone know whether Cliff Olson details what his process is? Obviously not all REM is equivalent but it would be interesting to get more details.

 

@user490 said:

Are the fitting targets generated by the hearing aid software or also by the REM software?

Both. It is the target that matters with manual real-ear measurement because it is what you're matching to. For automated real-ear measurement you need to enter the measured into the fitting software as this is what will actually control the REM equipment.

I'd be surprised if Dr Cliff measured RECDs as he hasn't worked with a paediatric population and most adult guidelines don't require it. Indeed, many clinics working with adults that have REM equipment lack the equipment to measure RECDs as it is available only at additional cost for non-Verifit equipment if you're not also purchasing a test box.

In reply to @Hearing_Potential

@Hearing_Potential said:

The same isn't required for adults as their ear canal volume generally remains stable so monitoring for sensorineural changes and comparing in HL isn't as big of an issue provided you're using the same insertphone insertion depth each time.

We only work with Adults and the way we test is Legally acceptable here. I’ll look to see if there’s a BAA reference for what you’re claiming, but there’s acceptable testing variation within the standard.

Also from my experience, the volumetric/positional change in testing only really appears to significantly affect the low frequency results vs Cans. If you know, you know: but given most don’t amplify below 1KHz, the point is largely moot for a large number of initial fits. I can see the issue being much more problematic in small canals with occluded fits though.

In reply to @Hearing_Potential

Think Cliff mentions test box so his best practices folks should have the equipment. Perhaps open a dialogue with him as it seems a conclusion from this may be that his comprehensive best practices are merely recommended/required and not exceeding in this respect so as to be best. A study could assess this so it seems someone would attempt that.

Pediatric patients report difficulties in transitioning to adult providers. There's discussion here attributing that to providers switching away from using a pediatric fitting formula. I wonder whether not measuring may be another factor patients can detect?

For this use, which RECD values do you input to the fitting software: 0.4cc WRECD, -1 RECD, Avg WRECD, Avg HA-1 RECD? Should I have my audiologist take a second measurement to confirm a match for the first as these don't change much in adults? I have no way to know whether the data is garbage in (I suppose I could first fit with and without measured RECD to see how it sounds but that may not be sensitive enough).

@user490 said:

For this use, which RECD values do you input to the fitting software: 0.4cc WRECD, HA-1 RECD, Avg WRECD, Avg HA-1 RECD?

You technically don't need to input an into the fitting software, only the verification equipment. Different hearing aid manufacturers deal with inputted RECD values differently and like first fit calculations they can do strange things when calculating gain.

I tend to only import RECD data into fitting software if it has been directly imported from the equipment into the Noah database as it provides the fitting software with more data points than you can typically enter manually. I don't do this for all manufacturers either.

The RECD type will depend on the equipment being used, but typically we're using -1 referenced coupler data with the fitting equipment converting from wideband RECD and HA-2 coupler measurements. It's something that requires a bit of training and experience to understand fully. I think most audiologists working solely with adults will consider it an unnecessary step.

There has been a single randomized control trial that included RECD measurement, but it was comparing the outcomes of a group of hearing aid users fitted with REM vs no REM. Not measured RECD + REM vs average RECD + REM. Most REM studies have failed to include RECD measurement as part of their protocols, because most hearing aids couldn't accurately hit prescription targets within less than a +/-5dB range on average when they were conducted back in the 1990s and early 2000s. Using an average measure didn't seem like much of a tradeoff at the time. Things have come a long way since then.

What do the versus average curves tend to look like? My -1 seems less wavy than examples I can find but it might be that WRECD is more wavy and HA-1 less so?

In reply to @Hearing_Potential

Would you say that running on an adult with a noticeably small canal would be important to an accurate target match with ? Also curious if you have any examples or can point me toward any examples. I'm very curious. Also makes me wonder about the calibration of the transducers used to measure the thresholds in the first place. Aren't they also impacted by this same effect?

Thank you!

@Um_bongo this article by Susan Scollie talks about the latest standard and addresses the key point about RECDs being used in the generation of prescription targets for real ear measurement.

In reply to @user490

They're unique to each individual ear. There can be common measurement errors you can detect, but every ear is unique and sometimes ears can be quite different in the same individual. Do you have personal access to a system?

In reply to @mindsil

I often find women with mild to moderate hearing losses often get hearing aids faster than men and I suspect some of this is because their loss is being measured using insertphones being placed in smaller ear canals, and their loss is greater than an equivalent hearing loss because of the way that insertphones are calibrated. It's important to correct for this in target generation by measuring ear specific rather than using an average to avoid overamplification in very small ear canals. The reverse is true for larger than normal ear canals.

In reply to @Hearing_Potential

No personal access to . Just tried to get an audiologist to take measurements for use with despite it not being as complete as REM.

The article posted states:

Regardless of the venting effects, earmold RECDs have more mid-frequency resonance and more roll off in the high frequencies, compared to foamtip RECDs. There is a substantial difference in opposite directions in the mid and the high frequencies between a foamtip RECD and an earmold RECD.

My RECD was measured using a mold (cShell) but the article is discussing molds which might be different with respect to RECD measurements. My loss is mid frequency so I wonder how much resonance might matter if the mold issue applies to both RIC and BTE molds?

The current standard is to use an -1 referenced (see the article I linked a few posts back). Hopefully that'll be what your audiologist used.

Edited

In reply to @Hearing_Potential

The main takeaway from that article seems to be how much doing an earmould based (with longer tubing) and applying that to /, produces considerable errors vs. foam inserts on short/calibrated tubing.

I get the need to understand testing variability, especially when there’s canal weirdness (we don’t test paeds). This is not applicable to our day to day and wouldn’t cause me to doubt our test procedures.

The other highly relevant point is that if you test into a given volume using inserts, to just before the second bend, then fit to the same canal to a similar position, then effect of residual volume changes (larger or smaller) will deviate in the same proportion and direction in both your fitting AND testing procedure. The method variation is self-limiting by nature.

It’s mainly through clinicians modifying the procedure, by introducing errors (testing through earmould/earmould tubing) that this ‘problem’ happens. Or testing through ‘supra-aural’ cans which don’t fit children’s bone structure/head size that you’ll likely encounter issues. I’m absolutely happy to accept that if your testing regime involves these issues then you should incorporate a modification to your fitting regime.

Duly noted that if we run into particularly small canals - paediatric insert size - I’ll keep an eye on the loudness, especially in low frequencies. Though like I noted, most manufacturers just wind this area down to avoid occlusive effects anyway.

In reply to @Hearing_Potential

@Hearing_Potential said:

I often find women with mild to moderate hearing losses often get hearing aids faster than men and I suspect some of this is because their loss is being measured using insertphones being placed in smaller ear canals, and their SPL loss is greater than an HL equivalent hearing loss because of the way that insertphones are calibrated.

That makes no sense: a smaller residual canal volume will give a both a better result and higher output from a similarly placed/occluded aid, which could lead to greater issues with occlusion/loudness rejection.

Correlation doesn’t equal causality. Women are often more socially pivotal and practical when men can prefer the ‘head in the sand’ approach to dealing with communication issues.

Edited
@Um_bongo said:

That makes no sense: a smaller residual canal volume will give a both a better Audiogram result

Think a little more. You get better because insertphones are calibrated for a larger volume than the average female ear canal so the sound pressure is louder than expected. But once the very positive is taken into account the hearing loss is in reality worse than the HL audiogram, hence greater listening difficulties for the same HL level of hearing loss.

Its the same reasoning behind why infant hearing losses measured in HL sometimes appear to worsen over time as they grow and we need to convert them back into SPL. The larger the cavity, the poorer the HL thresholds, therefore the smaller the cavity the better the HL thresholds. But the SPL thresholds will in reality be worse in a smaller ear vs a larger ear for the same HL thresholds when tested with insertphones due to the volume difference. How significant that difference is will vary between individuals. Measuring it is helpful, eliminating an unknown and reducing error.

I'm not claiming causation. I'm saying it is a possible contributing factor for why some women, particular those with smaller ear canals seem to adopt hearing aids at lesser degrees of HL loss than men, as a clinical observation from someone who is actually measuring RECDs for everyone.

In reply to @Um_bongo

@Um_bongo said:

The main takeaway from that article

The main takeaway I wanted you to pay attention to from the linked article was this paragraph:

"Sometimes you will hear clinicians say, “I use real ear. I don’t use ”.  In other words, they think if they are doing on-ear verification rather than coupler-based verification, then the RECD is not used.  This is a common misconception. In reality, the RECD may still be a factor behind the scenes in real ear verification, as shown in Figure 3.  If your thresholds were measured with insert earphones, then the RECD and the calibration values from the audiometer can be used to predict real ear thresholds. These will be displayed on the SPLogram. It is true, however, that direct verification of the real ear aided response () does not use the RECD at all.

So if you use real ear (good for you!), you will want to also measure the RECD so that it may be used behind the scenes to optimize the accuracy of the SPLogram and the prescriptive targets generated for the patient."

Related discussions