Exploring the utility of internal whistleblowing in healthcare via agent-based models

Abstract

Objective: The benefits of internal whistleblowing or speaking-up in the healthcare sector are significant. The a priori assumption that employee whistleblowing is always beneficial is, however, rarely examined. While recent research has begun to consider how the complex nature of healthcare institutions impact speaking-up rates, few have investigated the institutional processes and factors that facilitate or retard the benefits of speaking up. Here we consider how the efficacy of formal inquiries within organisations in response to employees' speaking up about their concerns affects the utility of internal whistleblowing.

Design: Using computational models, we consider how best to improve patient care through internal whistleblowing when resource and practical limitations constrain healthcare operation. We analyse the ramifications of varying organisational responses to employee concerns, given organisational and practical limitations.

Setting: Drawing on evidence from international research, we test the utility of whistleblowing policies in a variety of organisational settings. This includes institutions where whistleblowing inquiries are handled with varying rates of efficiency and accuracy.

Results: We find organisational inefficiencies can negatively impact the benefits of speaking up about bad patient care. We find that, given resource limitations and review inefficiencies, it can actually improve patient care if whistleblowing rates are limited. However, we demonstrate that including softer mechanisms for internal adjustment of healthcare practice (eg, peer to peer conversation) alongside whistleblowing policy can overcome these organisational limitations.

Conclusion: Healthcare organisations internationally have a variable record of responding to employees who speak up about their workplace concerns. Where organisations get this wrong, the consequences can be serious for patient care and staff well-being. The results of this study, therefore, have implications for researchers, policy makers and healthcare organisations internationally. We conclude with a call for further research on a more holistic understanding of the interplay between organisational structure and the benefits of whistleblowing to patient care.

Keywords: health quality improvement; internal whistleblowing; raising concerns; resource limitations; simulations; speaking-up.

Figure 1

Average patient care after 10 000 rounds.

Figure 2

Average patient care after 10 000 rounds when inquiry time is increased by 10% per additional review.

Figure 3

Percentage of good practice when a concern is resolved over 50 000 rounds for different whistleblowing rates and processing efficiencies. The chance to receive information about bad practice is held static at 100% (i=100). Solid lines: reviews initially take 100 rounds (model 2). Dotted lines: reviews take 50 rounds (model 3). Squares: no one ever blows the whistle, w=0. Circles, diamonds and triangles: w is 39%, 76% and 100%, respectively.

Figure 4

Patient care over 50 000 rounds when there is a chance e=0.01 that an individual misclassifies the valence of a practice. Solid lines: reviews initially take 100 rounds. Dotted lines: reviews take 50 rounds. Squares: no one ever blows the whistle, w=0. Circles, diamonds and triangles: w is 39%, 76% and 100%, respectively.

Figure 5

Patient care over 50 000 rounds when there is a e=0.01 chance a worker misclassifies a practice, and a a=0.001 * N that inquiries will leave bad practice unchanged. Solid lines: reviews initially take 100 rounds. Dotted lines: reviews take 50 rounds. Squares: no one ever blows the whistle, w=0. Circles, diamonds and triangles: w is 39%, 76% and 100%, respectively.

Figure 6

Comparison of patient care with and without soft advice. Patient care over 50 000 rounds where reviews initially require 100 rounds. i=1, e=0, a=0. Solid lines: no soft advice, s=0 (model 2). Dotted lines: soft advice; s=1. Squares: no one ever blows the whistle, w=0. Circles, diamonds and triangles: w is 39%, 76%, and 100%, respectively.