Hearing Loss in Percentages and Decibels [en]

As the founding editor of Phonak’s community blog “Open Ears” (now part of “Hearing Like Me“) I contributed a series of articles on hearing loss between 2014 and 2015. Here they are.

For years, I’ve been mystified when hearing people refer to their hearing loss in percentages. “I have lost 37% hearing in my left ear.”

Since I was thirteen and had my first audiogramme, that is how I’ve been thinking of hearing loss. In decibels, presented as a graph of how loud a sound needs to be so I can hear it, at various frequencies. I’ve showed my audiogramme on Open Ears already but here it is again:

Steph Audiogram

As you can see, at 500Hz I don’t hear sounds below 50dB, but at 4000Hz (higher pitch sounds) my left ear has almost “normal” hearing, as I can hear sounds as soft as 20dB. As is the case for most people, my hearing loss is not the same at all frequencies.

Hence the mystification: how could one express this with only one number? And how would you convert decibels (a logarithmic scale, where 20dB is 10 times as loud as 10dB) into percentages?

Many months ago already, I wanted to write a blog post about this. I did some research, asked some people, and stumbled upon a formula which didn’t feel very convincing (maybe this one). At the time already, it seemed to me that there was more than one method, which kind of discouraged me from further investigation.

A couple of months back, there was a consumer advocacy piece on Swiss TV where they sent somebody with hearing loss to various audiologists to see what solutions where recommended (and at what price). The surprising part was that the “hearing loss” of the test subject, expressed in percentages, varied wildly from shop to shop.

Anyway, this put me back on track to figure out how on earth they converted audiogrammes to percentages. Despite hearing people talk about their hearing loss in percentages all these years, I’d never been given a percentage value myself for mine.

I roped in Pascal to investigate and thanks to him finally got some satisfactory answers. Here are my take-aways.

First, the proper way to describe hearing loss is the audiogramme. As one can guess based on the results of the Swiss TV programme and the discussion around Christina’s article about “cheating” the test, taking somebody’s audiogramme is a bit of an art-form, although technically it is a rather simple procedure. Done well, it should produce the same result independently of who is measuring it (assuming your hearing is stable). This is, by the way, my personal experience with my audiogramme, done and redone over the years by three doctors and at least as many audiologists.

Second, there seems to be no end of formulas to “convert” audiogrammes to percentages, even in the United States alone (now extrapolate that to the rest of the world). And the results vary. According to one method, I have “25.3%” hearing loss. According to another, the one used by the doctors in the Swiss TV programme (CPT/AMA table reproduced below), I have “48.7%” in one ear and “40.7%” in the other.


Does this really mean anything? Does it make any sense to say I am missing “roughly half my hearing” or “a quarter of my hearing”? The first formula uses a kind of weighted average where you multiply the “good ear” by 5 — why on earth by 5? Quoting the article I just linked to: “Notably, while a five-to-one weighting is common among hearing impairment calculations, there is no research basis for this particular proportion.”

Third, again referring to the very interesting discussion in the same article, the need for a simple way to express hearing loss “objectively” seems to have its roots (at least some of them) in the compensations for work-related hearing loss. If you’re going to give money to a worker because of hearing lost on the job, there has to be an objective and simple way to determine how much. Which, when we realise that even an audiogramme is a rather poor indicator of the real-life impact of hearing loss. Two people with similar audiogrammes may feel differently impaired in their life by their hearing loss. Quoting again:

Few studies have found evidence for any of the several arithmetic hearing loss calculations in current or recent use in the US, as an effective measure of real-world hearing difficulty. More significantly, a literature review was unable to identify any study that has used appropriate statistical methods to evaluate the relative strength of association between these hearing impairment calculations and self-report measures.

Well, there we are. Measuring hearing loss is a hairy affair, and percentages don’t seem to me a very useful way of expressing it, as the calculation methods vary, seems sometimes pretty arbitrary, and apparently don’t correlate well with the real impact hearing loss has on our lives.

ADSL, bits et bytes [fr]

[en] Don't confuse bits and bytes like me. 1B (Byte) = 8b (bits). A 2000Kb/s DSL connection will only download 250KB/s, and that's normal.

Je viens d’apprendre quelque chose. Je sais que j’ai une ligne ADSL “2000/100”, ce que j’ai toujours traduit dans ma tête comme “Deux Mégas par seconde en download, et 100 K par seconde en upload.”

Que non. Je suis en train de transférer des données de mon serveur en Allemagne sur mon ordinateur. Mon serveur en Allemagne a une grosse bande passante (enfin, relativement: 100Mbits/s). Assez pour saturer une ligne ADSL, en principe. J’étais donc un peu surprise de constater que la vitesse de téléchargement que m’indiquait ma machine tournait autour de 250KB/s. J’ouvre une deuxième connection sur un autre serveur, et la vitesse de la première chute à 130KB/s — la deuxième, je vous le donne en mille, tournant autour des 120KB/s.

Mince, me dis-je. J’ai été trompée sur la marchandise! Eh oui. Je m’attendais en effet à voir un total autour de 2MB/s.

“Encore une victime du marketing,” me dit Patrick, avec qui j’avais partagé mon étonnement.

Une ligne de “2000”, c’est 2000Kbits/s. La taille des fichiers sur mon disque dur, par contre, est mesurée en KBytes. Bits. Bytes. Pas la même unité.

1B (Byte) = 8b (bits) *(pour se souvenir dans quel sens ça va, la grosse unité a droit au B majuscule)

Vous voyez où je veux en venir? On donne les vitesses des connections ADSL en bits, ça fait des nombres plus grands. 🙂 Du coup, cette confusion d’unités peut nous donner l’illusion que nos connections sont beaucoup plus rapides qu’elles ne le sont en réalité.

Donc, 2000Kbits/seconde, cela veut dire en fait environ 250KBytes/seconde (divisé par huit). Il faudra donc 4 secondes pour télécharger un fichier de 1MB (une grande photo par exemple, ou une minute de fichier mp3).

Alors voilà, ma connection ADSL marche très bien, mais elle est huit fois plus lente que ce que j’imaginais!