New CRNA vs. MDA and Patient outcome study

[gotosleep]

New Study Shows Surgical Death Rates Not Affected

by Type of Anesthesia Provider

Data reveals no significant differences in surgical mortality rates

when anesthesia is provided by nurse anesthetists

or anesthesiologists working individually or together.

PARK RIDGE, Illinois--Patients are just as safe receiving their anesthesia care from Certified Registered Nurse Anesthetists (CRNAs) or physician anesthesiologists working individually, or from CRNAs and anesthesiologists working in anesthesia care teams, according to a groundbreaking study published in the April 2003 AANA Journal.

The Institute of Medicine estimates that anesthesia care today is nearly 50 times safer than it was 20 years ago, with one anesthesia-related death per 200,000-300,000 cases. Despite this record of improvement, questions have remained about surgical patient safety related to types of anesthesia providers.

The study, titled "Surgical Mortality and Type of Anesthesia Provider," analyzed the effect of different types of anesthesia providers on the death rates of Medicare patients undergoing surgery. Researchers Michael Pine, MD, Kathleen Holt, PhD, and You-Bei Lou, PhD, studied 404,194 cases that took place from 1995-1997 in 22 states.

According to the study, surgical death rates were essentially the same whether anesthesiologists or nurse anesthetists provided the anesthesia individually or worked together in anesthesia care teams. Further, hospitals in which CRNAs were the only anesthesia providers had results similar to hospitals where anesthesiologists were involved in the anesthesia care.

Only cases with clear documentation of type of anesthesia provider were studied, and adjustments were made for differences in case mix, clinical risk factors, hospital characteristics, and geographic location. The types of surgical procedures included carotid endarterectomies, cholecystectomies, herniorrhaphies, mastectomies, hysterectomies, laminectomies, prostatectomies, and knee replacements.

"The results of this study are significant, particularly in this time of anesthesia provider shortages and rising healthcare costs," said Rodney C. Lester, CRNA, PhD, president of the 30,000-member American Association of Nurse Anesthetists (AANA). "It confirms what the AANA has been saying all along: Anesthesia today is safer than it has ever been, regardless of whether the anesthesia provider is a CRNA or an anesthesiologist.

"With the demand for surgical care and other procedures requiring anesthesia growing annually, and an insufficient number of qualified anesthesia providers to satisfy this demand, it is important that the current supply of CRNAs and anesthesiologists be used effectively," said Lester. "It should give patients great comfort to know that they are receiving the same high-quality anesthesia care whether it is provided by a CRNA or an anesthesiologist working individually or in a team."

Nurse anesthetists have been providing anesthesia care in the United States for more than 100 years. Today, CRNAs are the hands-on providers of approximately 65 percent of all U.S. anesthetics, and they are the primary anesthesia caregivers in the military, rural communities, and delivery rooms.

To read the study, visit the AANA Web site at http://www.aana.com and click on "Press Releases."

[Brenna's Dad]

haven't they already confirmed this sometime ago?

[Tenesma]

flawed study for several reasons:

1) it is retrospective - we need a prospective well balanced study

2) it only looks at mortality rates on mainly benign procedures - in which morbidity/length of stay etc would provide more information

3) "Medicare data do not distinguish between valid risk factors (ie, comorbid conditions) and inpatient complications, risk adjustment using Medicare data may fail to capture the true preoperative risk of death."

it also doesn't describe impact of pre-operative risk assessment, pre-operative medical management by anesthesia provider

the assumption that safety is determined by surgical death rate is insufficient...

but here is what really blows me away about this study which is kind of embarassing for the authors of that study: they were using predicted mortality rates based on old statistical analysis and didn't even compare actual mortality rates of the 400,000 cases they looked at.... what a waste of time, they could have done much better...

my 2 cents,

tenesma

PS here is link for those who want to read the study:

http://www.aana.com/press/2003/041103_pine.asp

[g8rlimey]

Thanks for your 2 cents, Tenesma. Are there such studies out there as you propose?

[yoga crna]

Tenesma, your comments reflect the fact that you are an M.D. anesthesiologist, something that you should disclose on your posts.

The CDC has stated many years ago that there is no necessity for a prospective study, because anesthesia mortality is very low and not a public health issue.

I think the authors of the paper are well regarded researchers who used the only non-biased information out there--Medicare data. I am not a researcher, but I do know that CRNA can and do provided competent care. If not, we would have been out of practice some time ago.

YogaCRNA (happy to disclose my credentials

[smiling_ru]

(1) it is retrospective - we need a prospective well balanced study.

I agree

2) it only looks at mortality rates on mainly benign procedures - in which morbidity/length of stay etc would provide more information.

Again I agree

3) "Medicare data do not distinguish between valid risk factors (ie, comorbid conditions) and inpatient complications, risk adjustment using Medicare data may fail to capture the true preoperative risk of death."

They did not use the medicare data to determine risk factors, for that very reason they used New Yorks SPARCS database

"Equations were derived to compute the probability of dying before discharge for each patient undergoing a procedure included in this study. Risk factors considered for each procedure were patients' age, sex, principal and secondary diagnoses, and selected information about procedures (eg, laparoscopic vs abdominal surgery). To identify potential risk factors, stepwise logistic regression8 was applied to New York's Statewide Planning and Research Cooperative System (SPARCS) data for 1996 and 1997. This state database was used exclusively to screen potential risk factors because, unlike the Medicare database, it distinguishes valid risk factors (ie, secondary diagnoses present on admission) from complications that occurred during hospitalization. Only comorbid conditions found to be statistically significant predictors of inpatient mortality (ie, P

(4) but here is what really blows me away about this study which is kind of embarassing for the authors of that study: they were using predicted mortality rates based on old statistical analysis and didn't even compare actual mortality rates of the 400,000 cases they looked at.... what a waste of time, they could have done much better...

Comparing actual mortality rates is exactly what they did, using the risk adjusted statistics to account for co-morbidities hospital type etc..

"Initial patient-specific mortality predictions were computed using procedure-specific risk-adjustment equations; predicted values then were adjusted using hospital-specific variables. The resulting predicted values were used to compare inpatient mortality rates among the 3 types of providers (anesthesiologists alone, CRNAs alone, anesthesia care teams)."

"For the 404,194 cases analyzed, Table 3 shows the distribution of patients among the 8 operations and the unadjusted mortality rate for each operation. Mortality rates ranged from 0.11% for mastectomies and for hysterectomies to 1.20% for cholecystectomies. The average for all patients was 0.38%. Anesthesia care was provided by anesthesiologists alone in 33.2% of cases, by CRNAs alone in 8.2%, and by anesthesia care teams in 58.6% (Table 4)."

Table 7 lists C statistics for the 8 clinical risk-adjustment models and for the final model incorporating clinical risk and hospital characteristics. Patient factors were most predictive of mortality for patients undergoing cholecystectomy or herniorrhaphy and least predictive for patients undergoing mastectomy or knee replacement

____________________________________________________

The truth is we do not know the mortality rates for any anesthesia provider either alone or combined. The 200,000 -300,00 number that is continually batted around is not valid. It was derived from a study in 1 hospital performed from 1975-1983, was quoted in the report from IOM To Err is Human, and has become the standard number we use.

We really do need some valid studies, but maybe we need to start with what is the risk in general regardless of provider type. Then move to specific provider types. It seems to me that the studies trying to show similarities or differences between provider types is a huge waste of resources, that could be better spent elsewhere. I have no doubt that if the studies are done they will show no difference in outcome, as this one did.

But, the political machine must be fed!

[meandragonbrett]

It's interesting that there was a MD involved with the study.

Brett

[Tenesma]

I felt after hanging out on your boards for this period of time that most would recognize me as an MDA (and the reason i hang out here every once in a while is because i enjoy all forums where there are active discussions about anesthetic management - that is why you don't see me interject on issues regarding CRNA training etc - my posts usually revolve around techniques and management - but this one caught my eye for sure)...

1) i agree with YogaCRNA that we don't need a study to prove anything since we all know that anesthesia has become a relatively low-mortality endeavor

2) but if we are going to do a study then lets do it right - hence prospective versus retrospective - and lets really make it interesting by comparing outcomes on ASA III/IV patients including morbidity (something that wasn't included in this "study")

3) Michael Pine is an MD/MBA who owns a consulting firm in Chicago that specializes in healthcare data - he has never been in an operating room

4) smiling-ru: you are right they used NY SPARC to help assess risk factors - but all of their data stems from medicare data which in turn... in fact i am just quoting their own discussion at the end of the article where the study says: "Medicare data do not distinguish between valid risk factors (ie, comorbid conditions) and inpatient complications, risk adjustment using Medicare data may fail to capture the true preoperative risk of death."

5) Smiling-ru: they never compared actual mortality data - if you read the study you will see that they used a whole bunch of complicated equations to help predict expected mortality rates for a set of operations... and that is all they did, everything else is silly extrapolation.... this study still has me scratching my head... they were trying to prove something without providing any data whatsoever.... now however, if you do choose to use their funky data you can take a look at their tables (linked from the article) where you can see that the procedure with the highest mortality were cholecystectomies, and the procedure with the highest percentage of independent CRNA care were cholecystectomies.... So if you are going to believe everything this study implies by silliness you would now have to also support their implied suggestion that maybe CRNAs provide a higher mortality risk for cholecystectomies

6) meandragonbrett: it is interesting that an MD was involved with this study... but you can find an MD to support any study, it is just unfortunate that this is such a poor study...

[Brenna's Dad]

I've only been able to skim over the article once, but felt I need to write at least a preliminary response.

1) I don't understand the significance of the statement, that the MDA involved in the study was not a clincian. It seems that much of the research being conducted in medicine is completed by non-clinicians, so I can't see why it should be of any significance in this case.

2) I believe that tone of Tenesma's post, is to imply that the sponsor's of the study (AANA) somehow bought the MD's findings. I get the impression, that she feels that if the MD was a 'real' clincian, that he wouldn't have supported the study.

(please correct me if I am wong).

3) I agree with the comment that the researchers never actually compared mortality rates for procedure based on specific provider. Rather strange.

The only data they do provide, is to compare the overall predicted mortality rate with the actual mortality rate based on independent provider or anesthesia care team. These are the statistics to which the ASA probably objects: that MDAs when working alone (according to the study) had a slightly higher overall actual mortality rate compared to predicted mortality rate (.41 vs. .39). CRNA's working alone had a slightly less overall mortality rate compared to predicted (.45 vs .50) and the anesthesia care team did better yet (.29 vs .35).

The procedure with the highest mortality rate (cholecystectomies) did indeed correlate with the procedure performed

most independently by CRNAs (13.07%). However, the second greatest independent procedure (10.04%) were prostatectomies, which had the third lowest mortality rate (.20%). The procedure performed most independently by MDAs were carotid endarterectomies (38.91%), which had the second highest mortality rate (.50%). The data provided just doesn't allow for accurate assumptions.

[smiling_ru]

Tenesma said 'Smiling-ru: they never compared actual mortality data - if you read the study you will see that they used a whole bunch of complicated equations to help predict expected mortality rates for a set of operations... and that is all they did, everything else is silly extrapolation.... this study still has me scratching my head... they were trying to prove something without providing any data whatsoever.... now however, if you do choose to use their funky data you can take a look at their tables (linked from the article) where you can see that the procedure with the highest mortality were cholecystectomies, and the procedure with the highest percentage of independent CRNA care were cholecystectomies.... So if you are going to believe everything this study implies by silliness you would now have to also support their implied suggestion that maybe CRNAs provide a higher mortality risk for cholecystectomies'

It seems to me that if you are comparing what the mortality should be (calculated from NY Sparcs) to what it is. Then that is evalutating the actual mortality data. What am I missing here?