Row the Amazon Cambridge Judge Business School logo Leonard Cheshire Disability logo

2077 miles. Two men. One boat.
Mark de Rond and Anton Wright will row the entire length of the Amazon River in September and October 2013.

Psychological safety on the Amazon

  • 12 AUG 2013
  • By Mark de Rond

Another week, another adventure. Our boat has arrived in Lima, Peru, but has been held for 'safeguarding' by our friends at customs. Fortunately help is at hand. Our fixer for Peru (formerly Chief of Police in Lima) assures us that, while customs officers keep changing their minds as to what to do with a golden yellow rowing boat, all will be well. Our Brazilian fixer, who once accompanied Sir Peter Blake on his fateful mission, has been active too. This email just in:

Dear Doc,

Regarding travel safety, I should warn you about the part of the river down bellow the Federal Police Control, known as base Anzol. It is particularly dangerous between Jutaí and Coari, where many bad occurred. You may find somebody at the Federal Police Control Base to give you some advises.

Otherwise my suggestion is to contract a speedboat with armed guards to follow you at a short distance giving you the necessary protection.

Please dont take any unecessary risk.

Physical safety can be insured against. Psychological safety cannot. Yet it is the latter, not former, that proves to be the most formidable obstacle to effective collaboration.

Amy Edmondson's work with hospital teams in the Boston area supplies some startling insights into the consequences of the psychological safety for patient welfare. Based on a sample of eight nursing units, she examined the link between, on the one hand, the quality of leadership and the strength of interpersonal relationships in nursing teams, and, on the other hand, errors in the administering of drugs to patients. Each team comprised forty staff on average, a combination of full- and part-time nurses, physicians, pharmacists, clerical and medical aids. Each was charged with a nurse manager. Edmondson collected survey data from the staff on various team characteristics, such as the amount of direction and coaching provided by nurse managers, relations between staff, and quality and style of leadership. As nurse managers were generally given a great deal of latitude to manage teams as they saw fit, there was plenty variation within the sample. Some teams were cohesive, tightly structured and carefully managed, others less so. Edmondson also collected data on medication errors from patient chart reviews and voluntary reports by team members. Her assumption was that better-managed teams were less likely to make mistakes in administering drugs than poorly managed teams.

A first glance at the collected data confirmed suspicions: the number of medication errors did vary substantially between the eight hospital nursing units, and were highly correlated with team characteristics. To her surprise, however, the sign was negative. Nursing units that might be textbook examples of tightly managed, well-structured, and well-managed teams were also those associated with the highest incidence of errors. Not just that, but this relationship only held for those mistakes that could have been avoided (rather than unexpected, drug-related complications over which the team wouldn't have had any control). And the difference was considerable: teams who rated their leadership highly, and who thought well of each other, showed ten times the number of drug administering errors. Intrigued, Edmondson and her colleagues decided to conduct a series of follow-up interviews with the staff, and also to observe them for short periods as they went about their everyday work. Doing so allowed her to get a more accurate sense of the varying social climates of these nursing units. When ranking the eight teams according to the openness of their climates, she discovered a near-perfect match between openness and medication errors, meaning that more errors meant a better reporting of mistakes. Teams that felt psychologically safe were those in which nurses, clerks, pharmacists and physicians were comfortable owning up to errors, whereas in those considered psychologically unsafe, mistakes went unreported. Armed with this information, you could be forgiven for picking the error-prone nursing unit over those ostensibly safer.

Her subsequent study of 16 cardiac surgery teams forced to implement a relatively new, minimally invasive, surgical procedure suggested that teams learn at different rates. Of particular interest is her observation that differences in the ability to surgical teams learn seems unrelated to variations in educational background and surgical experience. Nor are they related to the status of surgeons leading the different team, or levels of top management support for the new procedure. What mattered far more was the ability of surgeons to create a psychologically safe space for team members to admit mistakes, volunteer suggestions, and offer constructive criticism.

There is compelling evidence from experimentation in social psychology to suggest that when individual subjects are assigned to teams and given varying bits of information, they are far more likely to seek common ground than to explore differences in the information provided to each. Particularly when dealing with teams comprised of different areas of expertise, this natural tendency can greatly limit the ability of a team to learn or excel. The willingness to engage in cognitive conflict is imperative to team performance, but only ever happens in environments that are considered to be psychologically safe.

A related study in medicine looked at difficulties in medical teams as a result of unquestioned deference to expert status. Specifically, researchers tested the willingness of well-trained nurses to surrender their decision-making responsibilities to the judgment of what they thought was the attending physician. The experiment called for one of the researchers to make a call to 22 separate nurses' stations on various surgical, medical, pediatric, and psychiatric wards, identifying himself as a hospital physician and directing the answering nurse to give 20 milligrams of the drug Astrogen to a specific ward patient. In 95 per cent of the cases, the nurse went immediately to the ward medicine cabinet, secured the ordered dosage of the drug, and headed straight for the patient's room. This despite the fact that the drug had not been cleared for hospital use, the prescribed dosage was twice the maximum daily dose set by the manufacturer, and the directive was given by a person the nurse had never met or even talked with before on the phone. The authors of the study concluded that in fully staffed medical units like the ones they examined, it is natural to assume that multiple 'professional intelligences' cooperate to ensure that the best decisions are made. But in fact, under the conditions of the study, only one of those intelligences - the physicians' - may be functioning.

Corporate boards are not immune to psychological un-safety either. Jim Westphal and Michael Bednar's analysis of 228 boards of directors of medium-sized US companies provides compelling evidence for lack of psychological safety in the corporate world, using a much larger sample. Why is it, they asked, that boards so often fail to initiate strategic changes when faced with poor corporate performance? Prior research on executive decisions making has highlighted several cognitive biases associated with poor performance, for example, the tendency to over-attribute poor performance to circumstances outside of managerial control. Failure to change may also be due to a reluctance to admit that the current strategy isn't working, or even a conviction that with time the tide will turn provided one stays the course. There is even evidence to suggest that poor performance leads executives to restrict their search for new information and, particularly, data that casts doubt on the effectiveness of their current strategy. Occasionally this leads them to look outside for professional advice to help affirm their assumptions and bolsters their confidence. One of the reasons for having outside directors is for them to play devil's advocate and challenge explanations for poor corporate performance. That this may work better in theory than practice is exactly Westphal and Bednar's point. Their explanation for passivity in the face of poor performance is an interesting one. While outside directors may be more objective and less biased when it comes to evaluating corporate strategy, they are prone to 'pluralistic ignorance'. Their results provide strong evidence of the failure of board members to express their concerns with the status quo. What's more, they greatly underestimated the extent to which their fellow directors shared these concerns, with the predictable result that underperforming companies were more likely to persist with a doomed course of action. Popular examples of boards' failure to challenge corporate decisions include Royal Bank of Scotland's acquisition of ABN, and Lloyds' merger with Halifax.

So what can leaders do to boost the level of psychological safety within the teams they manage? One approach is to follow the example of the Royal Navy by 'de-ranking' when full and frank disclosure is required. A similar effect is achieved by Royal Navy chaplains who by virtue of their role will always assume the rank of whoever it is they happen to be talking to. It is a protocol not shared by either the Royal Army or Royal Air Force, nor particularly evident in corporate life.

It is on the Amazon. It's got to be.