This study suggests that residents can use QI methods to improve the quality of obesity screening in their clinics. The rates of height, weight, and BMI documentation significantly increased after the intervention. Even though they were not sustained after 1 year, rates of documentation were significantly higher than at baseline. Documentation of BMI was associated with higher rates of counseling obese patients about diet and exercise. Most notably, the process improvement of height measurement by nursing staff was sustained. Other changes, such as posting BMI charts in clinic areas, proved difficult to maintain.
There are many possible explanations for why rates of BMI documentation increased but were not sustained. One possible reason is because screening for obesity was perceived as futile by many residents because effective and feasible treatments for obesity do not exist. A similar cynicism toward obesity treatment has been expressed among the larger medical community,28
highlighting the need for more research on obesity treatments. It is worth noting, however, that the first-year residents, who were not present for the initial launch, still screened for obesity at higher rates after the QI project compared with baseline rates, which demonstrates the effectiveness of some system improvements. The decline in rates after 1 year also suggests that certain aspects of the intervention, such as obesity education and marketing of the QI project, may need to be repeated. Unlike BMI documentation, the rate of height measurement after 1 year was sustained, highlighting that nurses continued with the system improvement. This finding may be a result of differences in clinical practice between physicians and nurses. Because nurses generally have less autonomy in making clinical decisions than physicians,29
they may have adopted height measurement more easily as a part of their job responsibilities. Although relying on nurses instead of physicians to calculate BMI could lead to higher rates of obesity screening, it is unclear whether this would lead to appropriate counseling and treatment of obesity by physicians. Unfortunately, there was a low rate of dietitian referral for obese patients, even though rates of resident physician-led counseling increased. This may reflect the lack of accessible dietitian resources available in the resident clinic. Additionally, BMI charts were surprisingly difficult to retain, which may be because of high foot traffic and many health care workers in the patient rooms and clinic dictation areas.
This study has implications for involving residents in QI and for efforts to improve obesity screening. There were many challenges to implementing this resident-led QI initiative. Because residency is a period of training and career development, residents have many competing responsibilities including inpatient rotations, academic research, outpatient clinic, and education. These responsibilities make it challenging for residents to dedicate time to additional projects outside of requirements. In addition, resident turnover may limit the success of QI projects. Similar to other studies on QI projects involving residents,7,15,20
this study was able to show that a resident-led QI project could be associated with improved patient care. However, unlike previous studies, which have been conducted in family medicine15,18
residency programs and used a stand-alone PBLI curriculum7
or a task force,19
this study integrated the PBLI curriculum into an IM ambulatory curriculum10
with a team-based resident-led QI project. Additionally, the curriculum used in this study has been shown to improve resident knowledge about QI,30
just as many other studies on PBLI curricula.12,16,31
For example, 89% of residents rated their comfort level with PDSA cycles moderate to high after this curriculum, compared with only 9% of residents prior to the curriculum.30
This study adds to a very small literature about methods to improve obesity screening. In the only study on BMI documentation in which BMI charts were posted throughout the clinic areas, rates of BMI documentation were 49% in the intervention group compared with 21% and 32% in the control groups.32
Of note, our 1-year results are remarkably similar to the prior study, highlighting the importance of redesigned processes (ie, posting BMI charts, nurses measuring height, redesigned forms).
This study has several limitations. Although it was performed at a single academic clinic, it is likely that residents at other institutions will be able to adopt similar QI methods to improve clinical practice. Also, because of the constraints of resident time and turnover, data collection could only be cross-sectional and based on convenience samples. Therefore, conclusions regarding causality are limited.
In conclusion, our study suggests that residents can effectively use QI methods to improve the quality of patient care in residency clinics. This resident-led QI initiative resulted in sustained moderate improvements in obesity screening in a resident clinic. Future studies could evaluate if other resident-led QI projects can similarly improve patient care in resident clinics and if increased obesity screening is associated with improved patient outcomes.