The fact that a one-night PSG titration could very well be suspect (bankrupt is a bit severe I think) is probably the reason my sleep doctor always prescribes an APAP, and if my case is typical, puts the unit into APAP mode, has the patient come back after one month with the machine in tow, looks at the data, and sets a CPAP pressure.-SWS wrote:idamtnboy wrote:I'm concerned about making changes night to night. Doesn't give time for things to settle down.
Well, if we accept that premise as a rule rather than exception, then PSG titrations are a bankrupt proposition as a rule rather than exception. I agree that multiple nights at one setting can help with data interpretation amidst variability.
Judging from many comments from users here in the forum I think PSG titrations are often suspect. But, on the other hand, a good share of us on this forum are anal retentive about leaks, AHI, pressure, and the whole thing about CPAP therapy to a far greater degree than may be medically and statistically justifiable!
In the manufacturing world there is the management tool known as process control, measured by statistical process control methods. The underlying premise is that in the real and natural world of manufacturing all processes have some degree of variability inherent in them. There is no such thing as absolute. The control system relies on measuring the process for x consecutive number of times every nth increment of quantity or time. Typical numbers might be 5 times for every 100 pieces. This reliably captures the variability from piece to piece, and slow changes over time in the overall process. The average and spread of a data group is compared from group to group. I argue that the same approach can be used to measure physiological symptoms like blood glucose levels, AHI, and who knows what all. I am now checking my BG level every day for one week every six weeks. I will probably start doing that with my sleep parameters before long. The sleep doctor who says a one night PSG titration is accurate is like some engineers I knew one time. They had the attitude, "A single data point the universe describes." Ridiculous.
SU, sorry about hijacking your thread!





