There are other ways to assess
Ever heard of inter-rater reliability? It's a way of examining the differences between a number of people who are looking at the same thing. The problem with these "other ways" is that the inter-rater reliability is generally crappy mostly because of the fact that not everyone who looks hypoxic is hypoxic and some people have a better ability to separate the two. The reliability of these "other ways" in other words, is marginal at best. Since reliability seems to be a primary concern of yours, then I am surprised that you are not aware of this.
It is a good tool but, isn't an absolutely necessary tool and in many instances, not my most reliable tool.
It's either a good tool, or it's not reliable. Which is it? It is either a good tool or you can't rely on it a lot of the time. It sounds like you're basing this stance off of opinion and personal experience rather than anything valid or defensible.
Just in case you're interested:
Aughey K, Hess D, Eitel D, Bleecher K, Cooley M, Ogden C, Sabulsky N: An evaluation of pulse oximetry in prehospital care. Ann Emerg Med. 1991 Aug;20(8):887-91.
STUDY OBJECTIVES: We performed this study to evaluate the accuracy of pulse oximetry oxygen saturation (SpO2) against direct measurements of arterial oxygen saturation (SaO2) in the field. DESIGN: Prospective, cross-sectional, paired measurements of SpO2 against SaO2. SETTING: This evaluation was done in the prehospital setting. INTERVENTIONS: A pulse oximeter with digital probe was used to measure SpO2 in 30 patients. Arterial blood gases were drawn in the field while the pulse oximeter was in use, and oxygen saturation (HbO2) was measured by CO-oximetry. MAIN RESULTS: There was no significant difference between SpO2 (94.6 +/- 5.4%) and HbO2 (94.9 +/- 5.1%) (P = .495, beta less than .2). There was a strong correlation between SpO2 and HbO2 (r = .898). The bias between SpO2 and HbO2 was -0.3, with a precision of 2.4. When SpO2 was 88% or more, HbO2 was 90% or more in every case. Mean carboxyhemoglobin was 1.3 +/- 0.9%, and mean methemoglobin was 0.9 +/- 0.3%. There was no significant difference between the pulse oximeter heart rate and the ECG heart rate (P = .223, beta less than .2). CONCLUSION:
We conclude that pulse oximetry is sufficiently accurate to be useful in the field when SpO2 is more than 88%. It is potentially useful in patients with clinical signs of acute hypoxemia and in patients receiving interventions that may produce acute hypoxemia. Further work is needed to evaluate the accuracy of pulse oximetry in the settings of elevated carboxyhemoglobin, methemoglobin, and very low saturations.
BTW, "very low sats" is really a moot point in the field. If you have clinically correlatable reading that is less than 80% (most pulse oxs state not to rely upon a reading lower than 70-75%) you are not worry any further. It is quite overt hypoxemia and should be treated as such.
It always surprises me how few people know that pulse oximeters need to be calibrated.
Wu, a lot of modern pulse oximeters have a built in calibration check that automatically maintains the validity of the results. In fact, I believe it's effectively an industry standard nowadays. This is per a conversation I had with an engineer from Masimo (one of the major manufacturers of pulse oxs) where I asked about exactly what you just brought up.