5 principles for taking action in the face of uncertainty

While there are known prin­ciples for man­aging situ­ations of rel­at­ive uncer­tainty, it is less clear how to deal with situ­ations of great uncer­tainty, i.e. where the unknown reigns supreme. In these rare cases, there is no longer any ref­er­ence to exist­ing know­ledge (these are known as ‘unknown unknown’ situ­ations, as opposed to ‘known unknown’ situ­ations where there is an exist­ing ref­er­ence). In such cases, man­age­ment requires rules to those already used in the past.

The first wave of the COVID-19 crisis was an exem­plary case of this type of situ­ation¹. Dur­ing this peri­od, it was impossible to refer to a past event. Dubbed the ‘flu bug’ for a short time, the high mor­tal­ity rates observed quickly con­tra­dicted this asser­tion. What’s more, hos­pit­al stays in intens­ive care were much longer than those usu­ally observed for oth­er infec­tious res­pir­at­ory vir­uses, and loss of smell was a pre­vi­ously unknown symp­tom. There were innu­mer­able ‘sur­prises’, mak­ing any reas­on­ing by ref­er­ence to exist­ing know­ledge tricky.

In response, organ­isa­tions grappled their way through the pro­cess and made, often rushed, decisions. Some con­sidered that these actions were based on intu­ition and flair. How­ever, on closer inspec­tion, they reveal a cer­tain form of man­age­ment. This is shown by a study based on inter­views with more than 120 play­ers in the French hos­pit­al sys­tem, pub­lished recently in the European Man­age­ment Review².

The man­age­ment in ques­tion is based on a ration­al­ity that is neither Cartesian nor close to the concept of High-Reli­ab­il­ity-Organ­isa­tion (HRO), which is often used in uncer­tain situ­ations. Rather, it is a prag­mat­ic ration­al­ity, which con­sists of con­duct­ing a col­lect­ive invest­ig­a­tion to test hypo­theses in the field. Five prin­ciples emerge, reflect­ing the abil­ity of French hos­pit­al play­ers to demon­strate real­ism in the face of the unknown.

This first prin­ciple encour­ages us to design actions without wait­ing for per­fect inform­a­tion. The actions envis­aged at this stage rep­res­ent hypo­theses. They are derived from the ini­tial res­ults of the sur­vey, which selects them on the basis of the fol­low­ing meth­ods: (i) learn­ing from unusu­al events that can be observed in the field (anom­alies such as the abnor­mally long dur­a­tion of stays in intens­ive care for patients with COVID-19, which is sur­pris­ing giv­en our know­ledge of the effects of infec­tious vir­uses of the res­pir­at­ory sys­tem); (ii) con­sol­id­at­ing the reli­ab­il­ity of the inform­a­tion gathered by tri­an­gu­lat­ing dif­fer­ent sources or by assess­ing the pro­file of the issuer of the alert (what some people call ‘epi­stem­ic vigil­ance’); (iii) and when the hypo­theses are con­tra­dict­ory, to organ­ise debates between the vari­ous stake­hold­ers in order to reach a col­lect­ive decision (as in the case of the recep­tion of the Chinese curves on the incid­ence of the pan­dem­ic, which led to sev­er­al med­ic­al spe­cial­ists and epi­demi­olo­gists being brought togeth­er to com­pare their points of view). In this activ­ity, any mod­el­ling effort is also use­ful, but it is not suf­fi­cient alone, because it can neither provide reli­able con­text (the rela­tion­ship between nation­al pro­jec­tions and a loc­al situ­ation), nor provide a suf­fi­ciently con­sist­ent pre­dic­tion, due to a vari­ety of new cri­ter­ia that dis­rupt the ‘mod­el’.

This prin­ciple requires the hypo­thes­is to be tested in the field, which is the only way to judge its rel­ev­ance. Such recourse to the field takes place des­pite the sur­round­ing ignor­ance, and con­di­tions for action that are rarely optim­al. The object­ives are guided by a search for evid­ence and learn­ing, while the imple­ment­a­tion con­sists of cir­cum­scrib­ing the test in vari­ous small stages, each of which is fol­lowed by a rap­id return. This incre­ment­al approach optim­ises the assess­ment of the rel­ev­ance of the action, wast­ing as little time as pos­sible and avoid­ing obvi­ous errors. One example at the start of the COVID-19 crisis was the pri­or­ity actions taken in the health hos­pit­al sec­tor (lim­it­ing patient vis­its to hos­pit­al, set­ting up quar­ant­ine meas­ures, staff pro­tec­tion pro­ced­ures, increas­ing the num­ber of intens­ive care beds, to name but the main ones).These meas­ures quickly proved their worth. At the same time, they revealed the weak­ness of actions taken in EPHAD estab­lish­ments, where many eld­erly and vul­ner­able patients were exposed to the risk of the vir­us. This obser­va­tion led to the rap­id exten­sion and rein­force­ment of actions in this sec­tor, as recom­men­ded in the fol­low­ing principle.

This prin­ciple ensures that organ­isa­tions trans­late the res­ults of field tests appro­pri­ately, adapt­ing their actions if neces­sary. Depend­ing on the feed­back and the col­lect­ive delib­er­a­tion that fol­lows, the play­ers main­tain the ini­tial hypo­thes­is or modi­fy or even refor­mu­late it. The mul­tidiscip­lin­ary nature of the col­lect­ive delib­er­a­tions and the open-minded­ness of the par­ti­cipants are essen­tial cri­ter­ia in this revi­sion. The more the delib­er­a­tion involves a vari­ety of expert view­points, the more likely it is that the update will be rel­ev­ant.  Sim­il­arly, the more the atti­tudes expressed accept the con­clu­sions of the delib­er­a­tion, the more likely it is that the chosen hypo­thes­is will be adopted.

This third prin­ciple con­cludes an over­all approach. Com­pris­ing of three stages: (i) defin­i­tion of a hypo­thes­is; (ii) field test­ing; (iii) updat­ing of the hypo­thes­is on the basis of the res­ults, it is char­ac­ter­ist­ic of an abduc­tion meth­od. We start with a hypo­thes­is, check its rel­ev­ance by glean­ing observ­able facts, and then deduce wheth­er it should be main­tained or replaced by anoth­er. This approach is even more likely to be effect­ive if the invest­ig­a­tion is car­ried out col­lect­ively in order to cap­ture as many clues as pos­sible. It is also largely depend­ent on the atti­tudes of the mem­bers involved, as the fol­low­ing prin­ciple makes clear. 

Accord­ing to this prin­ciple, play­ers are encour­aged to express their doubts and to show humil­ity (fal­lib­il­ism), because the know­ledge they have acquired is extremely fra­gile, sub­ject to the appear­ance of new observ­able facts. Sim­il­arly, they are urged to avoid sim­pli­fic­a­tion by dicho­tom­ies between yes and no (dual­ist pos­i­tions), as these gen­er­ally reduce the abil­ity to select and inter­pret clues. Without these two atti­tudes, there is a high risk of mak­ing erro­neous decisions and giv­ing them too much weight.

An example of the import­ance of this prin­ciple was the debate on hydroxy­chloroquine as a treat­ment for the vir­us. The fact that the hypo­thes­is of such a treat­ment was put for­ward was not in itself shock­ing. It was even jus­ti­fied in the light of what was known about the sub­ject. On the oth­er hand, the fail­ure to ques­tion it when the tri­als car­ried out had not pro­duced con­vin­cing res­ults is evid­ence of an overly assert­ive, dual­ist­ic pos­i­tion in favour of the ‘yes’ option. 

In this quest for appro­pri­ate atti­tudes, the envir­on­ment out­side the play­ers involved in the sur­vey plays an import­ant role, likely to trig­ger harm­ful pressures.

Unknown situ­ations must be man­aged by those who have the most pre­cise know­ledge of the event. In par­tic­u­lar, if those on the ground have the expert­ise that is being built up (which is often the case), they must be giv­en pri­or­ity in the actions to be taken. One con­sequence of this is that they need to be pro­tec­ted from extern­al pres­sures out­side the scope of the invest­ig­a­tion. The lat­ter can in fact hamper the effort under­taken. The COVID-19 crisis high­lighted two such pressures:

1. Insti­tu­tion­al pres­sure, which may have called into ques­tion loc­al actions in the name of giv­ing pri­or­ity to decisions from high­er hier­arch­ic­al levels, even though these are irrel­ev­ant in terms of their respect­ive expertise.
2. Pres­sure from the media, which may have steered debates towards con­front­a­tions devoid of nuance, because of the forms of expres­sion imposed. Once again, the debate on hydroxy­chloroquine is a good example: the exchanges on TV often turned into cari­ca­tured oppos­i­tions between the for and against, lock­ing the play­ers into pos­i­tions of defend­ing a point of view, far removed from the atti­tudes of fal­lib­il­ity and anti-dual­ism that are neces­sary, but also dam­aging. As a res­ult, people work­ing in the field have found them­selves faced with patients who want to under­go treat­ment at all costs, even though there is no evid­ence to sug­gest otherwise. 

Unknown situ­ations must be man­aged by those who have the most pre­cise know­ledge of the event. 

The pre­vent­ive meas­ures against these extern­al pres­sures are to be found at the level of col­lect­ive delib­er­a­tions organ­ised by the play­ers hold­ing the expert­ise. They need to be cau­tious about mak­ing overly simplist­ic state­ments about the envir­on­ment. They must also pro­tect them­selves from pres­sure by unit­ing col­lect­ively. How­ever, this organ­ised pro­tec­tion must not cut off the expert­ise from inform­a­tion pro­duced else­where. The prin­ciple of col­lect­ive invest­ig­a­tion means that clues can be gleaned. They must there­fore strike a care­ful bal­ance between pro­tec­tion and selec­tion of sur­round­ing information. 

The five prin­ciples are inspired by the the­or­et­ic­al approaches of prag­mat­ism, a North Amer­ic­an school of thought from the early 20^th Cen­tury. Taken togeth­er, they form a spe­cif­ic form of man­age­ment at hos­pit­al level, known as High-Prag­mat­ic-Organ­isa­tion (HPO), in ref­er­ence to this school of thought. An HPO organ­ises col­lect­ive sur­veys, engages in abduc­tion pro­cesses, relies on play­ers whose atti­tudes cul­tiv­ate doubt and humil­ity, and pro­tects itself from extern­al pres­sure³⁴. By act­ing in this way, French hos­pit­al stake­hold­ers have shown that hos­pit­als can oper­ate accord­ing to their own prin­ciples to face up to the unknown.

This new con­cep­tu­al term also alludes to anoth­er concept, that of High-Reli­ab­il­ity-Organ­isa­tion (HRO), which is often used to express the man­age­ment required in situ­ations of uncer­tainty⁵. An HRO is an organ­isa­tion cap­able of deal­ing with crisis situ­ations where uncer­tainty reigns, by apply­ing dif­fer­ent prin­ciples (a hos­pit­al, but also a nuc­le­ar power plant or a sys­tem for organ­ising air flights, can thus be assim­il­ated to HROs in the event of a crisis).

Without going into detail, the pro­posed prin­ciples dif­fer from those just described on one essen­tial point: the absence of a ref­er­ence, and there­fore of the pos­sib­il­ity of express­ing reli­ab­il­ity. As a remind­er, reli­ab­il­ity aims to min­im­ise devi­ations from a nor­mal state that sets per­form­ance stand­ards⁶. It refers to anti­cip­a­tion by trig­ger­ing pre­vent­ive actions⁷. With HROs, nuc­le­ar power plant con­trol oper­at­ors can, for example, shut down react­ors if they think that oper­a­tions have entered these zones. Sim­il­arly, aero­plane pilots can refuse to fly if they think the equip­ment or weath­er con­di­tions are dangerous.

How­ever, when the play­ers are faced with the unknown, this reas­on­ing is inop­er­at­ive, because the very pur­pose of action is to define what these stand­ards are, and by deduc­tion the prin­ciples of pre­ven­tion. For example, when a remote mon­it­or­ing sys­tem was set up for patients with mild forms of COVID-19, the ini­tial feed­back showed just how rel­ev­ant the sys­tem was. Pre­vi­ously, man­agers had admit­ted these patients without real­ising that they were occupy­ing beds unne­ces­sar­ily and increas­ing the risk of the vir­us spread­ing. The pre­vent­ive nature of the action could only be judged fol­low­ing the empir­ic­al test.

Finally, it can be said that although the prin­ciples of ORH man­age­ment can be found, the absence of a ref­er­ence to nor­mal­ity lim­its the applic­a­tion of the concept in the event of an unknown situ­ation. In these situ­ations, the ref­er­ence is not determ­ined in advance but is con­struc­ted a pos­teri­ori. For this reas­on, the concept of HPO, and its five prin­ciples, seems more appro­pri­ate to deal with them.