User login
Burnout among health care professionals has been associated with lower quality of care, but the effect may be smaller than it seems, based on data from a meta-analysis of more than 200,000 clinicians.
Previous studies have reported associations between burnout and lower quality of care, but a standardized approach to analyze bias in the studies is lacking, wrote Daniel S. Tawfik, MD, of Stanford (Calif.) University and colleagues.
In a study published in the Annals of Internal Medicine, the researchers identified 123 publications from 1994 to 2019 with 142 study populations that included 241,553 health care providers.
Emotional exhaustion was the primary predictor for lower quality of care in 75 study populations, and overall burnout and depersonalization were the primary predictors for 56 and 11 study populations, respectively.
In an analysis of 114 unique burnout-quality combinations, 58 showed effects of burnout related to poor-quality care, 6 showed burnout related to high-quality care, and 50 showed no significant effect. Approximately one-third (33%) of the burnout-quality combinations were reported at least three times. In a review of the 46 burnout-quality combinations with primary effect sizes, 24 showed a significant effect of burnout on poor quality of care, 1 showed a significant effect of burnout on high quality of care, and 21 showed no significant effect.
The researchers also tested study bias using the Ioannidis test and found “an excess of observed versus predicted statistically significant studies (73% observed vs. 62%).”
The findings were limited by several factors, including the use of many cross-sectional, observational studies that could not show causality, the researchers noted. However, the results suggest several implications for future research including the need to consider exaggerated effects and reduce bias.
“Although the effect sizes in the published literature are modestly strong, our finding of excess significance implies that the true magnitude may be smaller than reported, and the studies that attempted to lower the risk of bias demonstrate fewer significant associations than the full evidence base,” the researchers noted.
“Whether curtailing burnout improves quality of care, or whether improving quality of care reduces burnout, is not yet known, and adequately powered and designed randomized trials will be indispensable in answering these questions,” they concluded.
The study was supported by the Stanford Maternal and Child Health Research Institute. Dr. Tawfik disclosed grants from Stanford Maternal and Child Health Research Institute during the study period.
SOURCE: Tawfik DS et al. Ann Intern Med. 2019 Oct 8. doi: 10.7326/M19-1152.
The current meta-analysis is consistent with previous research, but offers nothing new on the relationship between clinician burnout and quality of care, wrote Carolyn S. Dewa, MPH, PhD, Karen Nieuwenhuijsen, PhD, and Jeffrey S. Hoch, PhD, in an accompanying editorial.
Some of the concerns they expressed included variability in the methods used to measure provider burnout, as well as variability in measuring and defining medical error. They suggested that the researchers could have conducted a subgroup analysis based on error definition. “Such analyses might shed light on the types of errors associated with burnout and suggest directions for the design of robust psychometric studies about the error metrics,” they wrote.
The editorialists also expressed concerns about the heterogeneity of the studies included in the review and the potential for confounding. Finally, they noted that the use of observational studies in a meta-analysis can be challenging because “the assessment of observational studies is not straightforward.” They added that knowing the limitations of the studies is important in allowing readers to be confident in the estimates from any meta-analyses.
“Considering the limitations of the available literature, prior reviews, and Tawfik and colleagues’ current meta-analysis, we conclude that higher burnout is associated with lower quality, but we are left without clear answers about the magnitude or clinical significance of the relationship,” they wrote (Ann Intern Med. 2019 Oct 8. doi: 10.7326/M19-2760).
Dr. Dewa and Dr. Hoch are affiliated with the University of California, Davis. Dr. Nieuwenhuijsen is affiliated with the University of Amsterdam. The editorialists had no financial conflicts to disclose.
The current meta-analysis is consistent with previous research, but offers nothing new on the relationship between clinician burnout and quality of care, wrote Carolyn S. Dewa, MPH, PhD, Karen Nieuwenhuijsen, PhD, and Jeffrey S. Hoch, PhD, in an accompanying editorial.
Some of the concerns they expressed included variability in the methods used to measure provider burnout, as well as variability in measuring and defining medical error. They suggested that the researchers could have conducted a subgroup analysis based on error definition. “Such analyses might shed light on the types of errors associated with burnout and suggest directions for the design of robust psychometric studies about the error metrics,” they wrote.
The editorialists also expressed concerns about the heterogeneity of the studies included in the review and the potential for confounding. Finally, they noted that the use of observational studies in a meta-analysis can be challenging because “the assessment of observational studies is not straightforward.” They added that knowing the limitations of the studies is important in allowing readers to be confident in the estimates from any meta-analyses.
“Considering the limitations of the available literature, prior reviews, and Tawfik and colleagues’ current meta-analysis, we conclude that higher burnout is associated with lower quality, but we are left without clear answers about the magnitude or clinical significance of the relationship,” they wrote (Ann Intern Med. 2019 Oct 8. doi: 10.7326/M19-2760).
Dr. Dewa and Dr. Hoch are affiliated with the University of California, Davis. Dr. Nieuwenhuijsen is affiliated with the University of Amsterdam. The editorialists had no financial conflicts to disclose.
The current meta-analysis is consistent with previous research, but offers nothing new on the relationship between clinician burnout and quality of care, wrote Carolyn S. Dewa, MPH, PhD, Karen Nieuwenhuijsen, PhD, and Jeffrey S. Hoch, PhD, in an accompanying editorial.
Some of the concerns they expressed included variability in the methods used to measure provider burnout, as well as variability in measuring and defining medical error. They suggested that the researchers could have conducted a subgroup analysis based on error definition. “Such analyses might shed light on the types of errors associated with burnout and suggest directions for the design of robust psychometric studies about the error metrics,” they wrote.
The editorialists also expressed concerns about the heterogeneity of the studies included in the review and the potential for confounding. Finally, they noted that the use of observational studies in a meta-analysis can be challenging because “the assessment of observational studies is not straightforward.” They added that knowing the limitations of the studies is important in allowing readers to be confident in the estimates from any meta-analyses.
“Considering the limitations of the available literature, prior reviews, and Tawfik and colleagues’ current meta-analysis, we conclude that higher burnout is associated with lower quality, but we are left without clear answers about the magnitude or clinical significance of the relationship,” they wrote (Ann Intern Med. 2019 Oct 8. doi: 10.7326/M19-2760).
Dr. Dewa and Dr. Hoch are affiliated with the University of California, Davis. Dr. Nieuwenhuijsen is affiliated with the University of Amsterdam. The editorialists had no financial conflicts to disclose.
Burnout among health care professionals has been associated with lower quality of care, but the effect may be smaller than it seems, based on data from a meta-analysis of more than 200,000 clinicians.
Previous studies have reported associations between burnout and lower quality of care, but a standardized approach to analyze bias in the studies is lacking, wrote Daniel S. Tawfik, MD, of Stanford (Calif.) University and colleagues.
In a study published in the Annals of Internal Medicine, the researchers identified 123 publications from 1994 to 2019 with 142 study populations that included 241,553 health care providers.
Emotional exhaustion was the primary predictor for lower quality of care in 75 study populations, and overall burnout and depersonalization were the primary predictors for 56 and 11 study populations, respectively.
In an analysis of 114 unique burnout-quality combinations, 58 showed effects of burnout related to poor-quality care, 6 showed burnout related to high-quality care, and 50 showed no significant effect. Approximately one-third (33%) of the burnout-quality combinations were reported at least three times. In a review of the 46 burnout-quality combinations with primary effect sizes, 24 showed a significant effect of burnout on poor quality of care, 1 showed a significant effect of burnout on high quality of care, and 21 showed no significant effect.
The researchers also tested study bias using the Ioannidis test and found “an excess of observed versus predicted statistically significant studies (73% observed vs. 62%).”
The findings were limited by several factors, including the use of many cross-sectional, observational studies that could not show causality, the researchers noted. However, the results suggest several implications for future research including the need to consider exaggerated effects and reduce bias.
“Although the effect sizes in the published literature are modestly strong, our finding of excess significance implies that the true magnitude may be smaller than reported, and the studies that attempted to lower the risk of bias demonstrate fewer significant associations than the full evidence base,” the researchers noted.
“Whether curtailing burnout improves quality of care, or whether improving quality of care reduces burnout, is not yet known, and adequately powered and designed randomized trials will be indispensable in answering these questions,” they concluded.
The study was supported by the Stanford Maternal and Child Health Research Institute. Dr. Tawfik disclosed grants from Stanford Maternal and Child Health Research Institute during the study period.
SOURCE: Tawfik DS et al. Ann Intern Med. 2019 Oct 8. doi: 10.7326/M19-1152.
Burnout among health care professionals has been associated with lower quality of care, but the effect may be smaller than it seems, based on data from a meta-analysis of more than 200,000 clinicians.
Previous studies have reported associations between burnout and lower quality of care, but a standardized approach to analyze bias in the studies is lacking, wrote Daniel S. Tawfik, MD, of Stanford (Calif.) University and colleagues.
In a study published in the Annals of Internal Medicine, the researchers identified 123 publications from 1994 to 2019 with 142 study populations that included 241,553 health care providers.
Emotional exhaustion was the primary predictor for lower quality of care in 75 study populations, and overall burnout and depersonalization were the primary predictors for 56 and 11 study populations, respectively.
In an analysis of 114 unique burnout-quality combinations, 58 showed effects of burnout related to poor-quality care, 6 showed burnout related to high-quality care, and 50 showed no significant effect. Approximately one-third (33%) of the burnout-quality combinations were reported at least three times. In a review of the 46 burnout-quality combinations with primary effect sizes, 24 showed a significant effect of burnout on poor quality of care, 1 showed a significant effect of burnout on high quality of care, and 21 showed no significant effect.
The researchers also tested study bias using the Ioannidis test and found “an excess of observed versus predicted statistically significant studies (73% observed vs. 62%).”
The findings were limited by several factors, including the use of many cross-sectional, observational studies that could not show causality, the researchers noted. However, the results suggest several implications for future research including the need to consider exaggerated effects and reduce bias.
“Although the effect sizes in the published literature are modestly strong, our finding of excess significance implies that the true magnitude may be smaller than reported, and the studies that attempted to lower the risk of bias demonstrate fewer significant associations than the full evidence base,” the researchers noted.
“Whether curtailing burnout improves quality of care, or whether improving quality of care reduces burnout, is not yet known, and adequately powered and designed randomized trials will be indispensable in answering these questions,” they concluded.
The study was supported by the Stanford Maternal and Child Health Research Institute. Dr. Tawfik disclosed grants from Stanford Maternal and Child Health Research Institute during the study period.
SOURCE: Tawfik DS et al. Ann Intern Med. 2019 Oct 8. doi: 10.7326/M19-1152.
FROM THE ANNALS OF INTERNAL MEDICINE
