Author: Michael Vlessides
Assessing patient frailty preoperatively is recommended by various practice guidelines, but which instrument should clinicians use?
A new study has concluded that the Clinical Frailty Scale is superior to both the Fried Phenotype and Frailty Index when assessing risk for a variety of negative postoperative outcomes.
“Frailty has emerged as an important perioperative risk factor,” said Emma Harris, MD, a resident at the University of Ottawa , in Ontario. “Recent systematic reviews have told us that frailty increases perioperative risk of mortality, complications and health care utilization. Furthermore, there are societal guidelines that recommend routine preoperative assessment of frailty. Nevertheless, most institutions still do not routinely assess frailty preoperatively.
“One of the potential barriers to this is that there are many different instruments out there and we don’t actually know which is the best to use,” Dr. Harris said.
To explore this question further, Dr. Harris and her colleagues compared three leading frailty instruments. Specifically, they measured the increases in accuracy achieved by adding each instrument to traditional patient preoperative risk factors when predicting adverse postoperative outcomes.
Wide Variation in Measuring Frailty
“The Clinical Frailty Scale uses a scale from 1 to 9, where each number is associated with a short vignette and an image,” Dr. Harris explained. “So you can assign what’s most appropriate to the patient in front of you.
“The Fried Phenotype, on the other hand, looks at frailty as a syndrome, something with signs and symptoms that can be measured,” she said. “In this case, you’re assigning 1 point each to things like walking speed and weight loss, for a final score between 0 and 5.”
The primary outcome of the analysis was death or new disability. Secondary outcomes included prolonged hospital length of stay. Accuracy measures were based on Steyerberg’s framework, using model fit, discrimination, calibration, reclassification and explained variance.
“We wanted to choose outcomes that patients have told us are important to them,” Dr. Harris said. “They told us they care about going home and maintaining their function after surgery.”
The researchers began with a baseline score composed of four patient risk factors: age, sex, ASA physical status class and procedural risk. Next, they assessed the ability of each of the three frailty instruments to bolster the predictive ability of the baseline score.
In total, 645 patients were included in the study cohort, for whom data were collected at three different centers between 2015 and 2017.
In presenting the study at the 2019 annual meeting of the Canadian Anesthesiologists’ Society (abstract 624396), Dr. Harris reported that 72 patients in the original cohort (11.2%) died or experienced a new disability. All of the tested indexes improved model fit, discrimination, overall reclassification, and explained variance for the baseline model with respect to the primary outcome of death or new disability.
The addition of the Frailty Index led to overprediction of risk in higher-risk patients and improperly reclassified individuals who experienced an event. Moreover, the addition of the Clinical Frailty Scale led to the greatest increase in discrimination and explained variance with respect to death and new disability.
Prolonged hospital length of stay occurred in 164 patients (25.4%). The study found that the Clinical Frailty Scale improved each accuracy measure as much as or more than the Fried Phenotype or Frailty Index. Furthermore, the Clinical Frailty Scale was the only instrument to improve reclassification of both events and nonevents.
Finally, 60 patients (9.3%) experienced an adverse discharge. The Clinical Frailty Scale was the only instrument found to improve model fit. In addition, the scale provided the greatest increase in discrimination and explained variance, and provided the best calibration across the risk spectrum.
“So, the take-home message here is I would recommend that if you assess frailty, you use the Clinical Frailty Scale,” Dr. Harris concluded. “We’ve shown it’s the most accurate, and in previous studies they’ve shown that it’s quick and easy to use in the preassessment clinic.”
The Personal Touch
Gregory Hare, MD, PhD, was impressed with the study. “This is terrific work, a really important study,” said Dr. Hare, a professor of anesthesia at the University of Toronto Faculty of Medicine. “Is it atypical or surprising that you’ve consistently shown that perhaps the simplest scale is the best?”
“The Frailty Index has 30 variables, so you would think that it has the most information and would therefore be the most useful,” Dr. Harris replied. “I think there is something to be said for the Clinical Frailty Scale, where you put some of your clinical acumen into it. You’re taking into account factors that you can’t necessarily put your finger on when you’re assigning a number on another scale.”
Alana Flexman, MD, also was surprised that the Clinical Frailty Scale proved superior. “I also would have thought that the Frailty Index would have performed better because it’s such a big scale,” she said, a clinical associate professor in anesthesiology at the University of British Columbia, in Vancouver. “I agree there’s a lot of intrinsic factors that we’re taking into account with the Clinical Frailty Scale.
“I imagine your assessors were trained and probably used the Clinical Frailty Scale quite a bit,” Dr. Flexman added. “How do you think that would apply to a population of anesthesiologists who don’t regularly use the scale? Do you think you’d get the same results if they weren’t trained as well?”
“The Clinical Frailty Scale has been validated with different raters,” Dr. Harris replied. “There’s actually a study being done right now with patients rating themselves, and early results have shown that works as well. That would suggest to me that you don’t need a whole lot of training to appropriately use it.”