-SWS wrote:I personally didn't find that fifth index to be inherently over-reaching. My take was that index five is a highly typical method of dismissing artifacts: if prior cycles were also found to be flow limited, then index five carries a low-probability weighting toward final, sum-total p; if prior cycles were normal, then index five carries a high-probability weighting toward that final, sum-total p (probabilities weighted toward 0, w/ 0=FL or weighted toward 1, w/1=non-FL).
So in my view index 1 through index 4 collectively assess four unique wave shape characteristics of a single inspiratory cycle, presently receiving calculation for a semi-final probability p. Then index five steps in with its context/temporal sensitive factorization toward weeding out artifacts.
To be honest, -SWS, after skimming through a bunch of very similar patents, most with similar claims, I suddenly keyed in on that fifth index because it just looked loosey-goosey. Since I didn't elaborate earlier, let me elaborate now.
Patent 5335654:
This is the fundamental patent. Regarding flow-limitations it says:
Patent 5335654 wrote:9. The apparatus of claim 7, wherein said signal processor further comprises means for determining the presence of patient airway obstruction when the contour of said waveform corresponding to inspiration includes a plateau.
10. The apparatus of claim 7, wherein said signal processor further comprises means for determining the presence of patient airway obstruction when a portion of the contour of said waveform corresponding to inspiration is flattened.
11. The apparatus of claim 7, wherein said signal processor further comprises means for determining the presence of patient airway obstruction when the contour of said waveform corresponding to inspiration deviates from a substantially sinusoidal shape.
These three claims are the only qualifying details provided which describe waveshape processing. Otherwise, the claims are very nebulous and extremely broad. Don't get me wrong though, Rapoport was crystal clear in the description of his invention about what he was proposing. In the claims he backed that up by stating his processing was looking for a "flattened" waveshape. So far, so good.
I see this morning that patent 5490502 is where Rapoport introduced the five indices which are front and center stage here, beginning with the very first claim:
Patent 5490502 wrote:What is claimed is:
1. A method for optimizing the positive airway pressure to a patient, comprising the steps of:
a) applying an initial level of positive airway pressure of a breathing gas to a patient;
b) detecting the onset of an inspiratory flow of breathing gas to the patient;
c) storing data values representative of the inspiratory flow of breathing gas to the patient;
d) determining whether the stored data values indicate a flow limitation in the patient; and
e) increasing the positive airway pressure when the stored data values indicate a flow limitation in the patient.
2. The method of claim 1, further comprising the step of decreasing the positive airway pressure when the stored data values do not indicate a flow limitation in the patient.
3. The method of claim 1, wherein step d) includes comparing the stored data values with a sinusoidal contour.
4. The method of claim 3, wherein step d) includes correlating the stored data values with a pure sine wave.
5. The method of claim 1, wherein step d) includes analyzing the stored data values for flatness.
6. The method of claim 5, wherein step d) includes comparing a regression fit of the stored data values with a regression fit of a pure sine wave.
7. The method of claim 1, wherein step d) includes analyzing the stored data values for respiratory effort.
8. The method of claim 7, wherein step d) includes comparing a peak value of the stored data values with a peak value of a derivative of the stored data values.
9. The method of claim 1, wherein step d) includes analyzing the stored data values for relative flow magnitude.
10. The method of claim 9, further comprising repeating steps b) through d) for a plurality of patient inspirations to create a plurality of peak flow values for flow limited breaths and to create a plurality of peak flow values for non-flow limited breaths, and comparing a peak value of the stored data values with an average of the plurality of peak flow values for flow limited breaths and with an average of the plurality of peak flow values for non-flow limited breaths.
11. The method of claim 1, further comprising the steps of:
f) repeating steps b) through d) for a plurality of patient inspirations to create a plurality of peak flow values for flow limited breaths and to create a plurality of peak flow values for non-flow limited breaths;
g) correlating the stored data values with a pure sine wave to create a first index;
h) comparing a regression fit of the stored data values with a regression fit of a pure sine wave to create a second index;
i) comparing a peak value of the stored data values with a peak value of a derivative of the stored data values to create a third index; and
j) comparing a peak value of the stored data values with an average of the plurality of peak flow values for flow limited breaths and with an average of the plurality of peak flow values for non-flow limited breaths to create a fourth index.
12. The method of claim 11, wherein step d) further includes the step of calculating a fifth index as a function of the first index, the second index, the third index and the fourth index, wherein the fifth index indicates the presence of flow limitation in the patient.
13. The method of claim 12, wherein step d) further includes the step of calculating a fifth index as a mathematical function of the first index, the second index, the third index and the fourth index, wherein each index includes a weighted coefficient having a range including a value of zero.
Patent 535739 adds nothing on flow limitation, but is an interesting patent which goes off in the direction of a mini sleep-lab, including a position sensor, oximeter, and expiratory airflow measurement. For example:
Patent 535739 wrote:2. The device of claim 1, further comprising a position sensor configured to generate third data values representative of the physical position of the patient, wherein said computer memory is further configured to store the third data values.
3. The breathing device of claim 1, further comprising an oximeter configured to generate fourth data values representative of the oxyhemoglobin saturation level of the blood of the patient, wherein said computer memory is further configured to store the fourth data values.
4. The device of claim 1, wherein said flow sensor is further configured to generate fifth data values representative of an expiratory flow of ambient air from the patient, and said computer memory is further configured to store the fifth data values.
Patent 5546933 simply adds expiratory flow to the mix:
Patent 5546933 wrote:What is claimed is:
1. A method for detecting flow limitations in the airway of a patient, comprising the steps of:
a) providing a nasal fitting in fluid communication with the airway of a patient, the nasal fitting having means for measuring the inspiratory flow of air to the patient;
b) measuring the inspiratory flow of air to the airway of the patient;
c) generating a curve of data values representative of the inspiratory flow of air to the patient for each breathing cycle;
d) processing such data values to generate a plurality of indices each independently descriptive of the shape of each such curve; and
e) combining such indices to determine whether a flow limitation in the patient is indicated.
2. The method of claim 1, further comprising the step of measuring the expiratory flow of ambient air from the patient, and generating second data values representative of the expiratory flow.
In Patent 5803066, Rapoport uses six indices by adding the area of the inspiratory waveform to the mix. It seems obvious here that Rapoport needs additional discriminators to accurately identify flow-limitation. No doubt flow-limitation detection has presented itself as a more difficult problem than he first imagined. This is also the patent where cardiac sensing was introduced to determine airway patency.
Patent 5803066 wrote:What is claimed is:
1. A method for optimizing the positive airway pressure to a patient, comprising the steps of:
a) applying an initial level of positive airway pressure of a breathing gas to a patient;
b) detecting the onset of an inspiratory flow of breathing gas to the patient;
c) storing data values representative of the inspiratory flow of breathing gas to the patient;
d) determining whether the stored data values indicate a flow limitation in the patient by calculating the ratio of the area of the inspiratory waveform to the area of a pure sine wave to create a first index; and
e) increasing the positive airway pressure when the stored data values indicate a flow limitation in the patient.
In Patent 6299581, Rapoport goes back to the mini sleep lab thing:
Patent 6299581 wrote:1. A method for detecting flow limitation in an airway of a patient breathing air in an unassisted manner, comprising the steps of:
...
2. The method of claim 1 wherein said patient is in a physical position, further comprising the steps of:
generating with a position sensor a third set of data values representative of the physical position of the patient; and
...
3. The method of claim 1 wherein blood circulating within said patient has an oxyhemoglobin saturation level, further comprising the steps of:
generating with an oximeter a fourth set of data values representative of the oxyhemoglobin saturation level of the blood of the patient ...
Patent 6793629 doesn't seem to add anything new to the flow-limitation arena. I'm not quite sure what the point of this patent was, although some of the wording seems to suggest a diagnostic slant rather than a treatment regimen.
-SWS wrote:But I do agree with your high-level assessment, Bill: I very strongly suspect the 420e carries a low specificity for FL detection. In other words I also happen to suspect the 420e is generating quite a few false positives for FL---much more so for certain patients possessing the right characteristics for FL false positive readings on the 420e.
Yeah, it's those false positives which kept Rapoport busy trying to find a solution.
-SWS wrote:Regardless, I think the 420e is accurately measuring something---perhaps a variety of conditions across the patient population---when it does accurately detect inspiratory wave shapes that significantly deviate from that sinusoidal reference. The question in my mind is how many different conditions in physiology might be capable of generating these FL false positives on the 420e? How many are benign? How many are SDB related? And how many are potentially problematic with respect to health?
Agreed! The 420E is measuring something. For many folks flow-limitation detection apparently works but, if ozij's poll is an indication, there are just as many folks for whom flow-limitation detection doesn't work, or at least not for the intended purpose.
Flow-limitation detection seems simple enough, doesn't it? Rapoport apparently thought so, too, back around 1992 when his patent was filed. His more recent patents which analyze breathing to determine the "patient's state" suggest that the complexities of breathing patterns have drawn him into considerably further analysis of this seemingly simple area.
Regards,
Bill (who still has a few comments to make regarding his own apnea in relation to GERD ... )
