The Hawthorne Effect Revisited

    loading  Checking for direct PDF access through Ovid

Excerpt

Manufacturing productivity studies done at the Western Electric plant in suburban Chicago (Hawthorne) in the late 1920s and early 1930s found that environmental changes (improved lighting) that were made at the facility improved productivity. Thirty years later, the results of this manufacturing experiment were reinterpreted by Landsberger1 to be the result of employee awareness that they were being observed in the performance of their work duties rather than environmental changes of the workplace. Improved performance because of awareness that performance was being monitored became known as the Hawthorne effect. In 1970, Cruse2 demonstrated that monitoring surgical outcomes with a surveillance nurse had a Hawthorne-like effect in reducing surgical site infection (SSI) rates. The act of performing a study or actively observing events in health care in all likelihood will independently change the outcome, especially when the outcome is being indexed to previous performance or to an external benchmark.
The latest testimonial to the Hawthorne effect in surgical outcomes can be observed in the study by Gorgun et al published in this issue of Diseases of the Colon & Rectum.3 The authors report on the implementation of 14 process measures that were used consistently by surgeons in a single facility to reduce SSIs after their facility had been identified by the National Surgical Quality Improvement Project (NSQIP) for having excessive infections. In this study, rates of SSI were evaluated and compared with a previous time interval. After implementation of the bundled process measures, they found that overall SSI rates declined to 6.6% from 11.8% in the prebundle period. Of particular interest, they identified that the most significant reduction occurred in organ-space infections from 5.5% to a rate of 1.7% after implementation of the bundle (p < 0.001).
On further review, one must be skeptical about the merits of the 14 process measures, because only the antibiotic bowel preparation has a potential benefit in the reduction of leaking anastomoses or intra-abdominal abscesses.4 Mechanical preparation of the colon alone has no benefit without the oral antibiotic bowel preparation,5 and the other 12 processes are only effective in the reduction of superficial SSIs. Many would argue that core body temperature control, supplemental oxygenation, glucose control, pulsed irrigation of the surgical site, topical antibiotics, and many other interventions should have been used. Indeed, there are ≥50 ways to cause SSIs in colon surgery and ≥50 methods for prevention.6 An argument can be made that the outcome improvement in this study comes from the awareness that performance is being observed and that improvements are expected. If performance is being observed and documented, then compliance with other processes that are not the focus of the study are likely to be included by the surgeon in patient management. It is yet another example of the Hawthorne effect being revisited.
Improvements in SSI rates are not the only adverse outcomes that need to be addressed in colorectal surgery today. Infectious and noninfectious complications continue to be common and result in considerable morbidity, mortality, and readmissions of patients. Perhaps more alarming than the overall rates of adverse outcomes in colorectal surgery are the dramatic differences in risk-adjusted outcomes among all hospitals.7,8 Deaths, major inpatient complications, and readmissions vary by 3-fold when 90 days of postdischarge outcomes are included in the evaluation. The differences in risk-adjusted outcomes between top and suboptimal performing hospitals indicate that considerable improvement is possible. My experience in evaluating and reporting hospital performance is that hospitals do not know their results of care, especially when the postdischarge period is included in the assessment. Hospitals currently depend too much on compliance, with process measures as the surrogate for measured outcomes.
    loading  Loading Related Articles