How to read an audiogram: symbols, degree, and PTA

An audiogram plots frequency on the horizontal axis, in Hz, increasing left to right on a logarithmic scale, typically from 125 or 250 Hz to 8000 Hz. Hearing level in dB HL runs on the vertical axis, increasing downward, typically from about -10 dB HL at the top to 110 or 120 dB HL at the bottom. The grid is drawn so one octave horizontally lines up with 20 dB vertically, which keeps slopes looking the way clinicians expect regardless of which two frequencies are compared.

The standard US symbol set, published by ASHA, uses a circle for unmasked air conduction on the right and an X on the left, connected across frequencies by a solid line. Masked air conduction switches to a triangle on the right and a square on the left. Bone conduction at the mastoid uses a less-than sign for an unmasked right ear and a greater-than sign for an unmasked left ear, switching to brackets when masked; bone symbols are conventionally not connected. Clinical convention colors the right ear red and the left blue, though ASHA notes color is not required to read the chart, since it is often lost when a chart is copied or faxed. An arrow angled outward and downward marks no response, and those symbols are never connected to neighboring points.

The classic three-frequency pure-tone average, sometimes called PTA512, is the mean of thresholds at 500, 1000, and 2000 Hz, the frequencies that matter most for understanding speech. It is the right number for degree of loss when the whole audiogram sits inside a single loss category, and it is what gets cited for hearing aid candidacy or as a comparison point elsewhere in the workup, such as checking agreement with the speech reception threshold.

Other averages exist for other purposes: a four-frequency average adding 3000 Hz appears in some asymmetry criteria, a different one adding 4000 Hz instead shows up in hearing aid work, and a two-frequency Fletcher average, the best two of 500, 1000, and 2000 Hz, suits a sharply sloping loss. State which frequencies were averaged, since more than one four-frequency convention exists.

Degree of loss is read off the three-frequency pure-tone average against a named scale. The Clark (1981) scale, the one ASHA itself publishes, runs normal from -10 to 15 dB HL, slight from 16 to 25, mild from 26 to 40, moderate from 41 to 55, moderately severe from 56 to 70, severe from 71 to 90, and profound from 91 dB HL and above. An older but still commonly taught scale, the Goodman scale, agrees with Clark from 41 dB HL upward but has no slight category, calling -10 to 26 dB HL within normal limits and 27 to 40 dB HL mild.

The two scales disagree only in that low range, making a threshold in the -10 to 26 dB HL band the one place a degree label is ambiguous unless the scale is named. Above 26 dB HL, both scales agree, so labels there are safe.

The pure-tone average only tells the whole story when every tested frequency falls in one category. A loss that is mild at 250 Hz and severe at 8000 Hz is not accurately called moderate just because those numbers average out that way. Describe a spanning loss as a range, worded in the direction it slopes: mild sloping to severe for a loss worsening as frequency rises, severe rising to mild for the reverse, and a plain mild to severe, no directional word, for a notch or trough shape. The PTA-based label stays useful elsewhere in the chart, but it answers a narrower question than the overall degree of a spanning loss.