A single cluster analysis was performed on the whole sample (Togo + France) using the K-means method advocated by Hofmans and Mullet (2013). A six-cluster solution was retained based on the technique advocated by Schepers and Hofmans (2009).

The patterns of data that correspond to each cluster are shown in Figure 1. In each panel, judged appropriateness of the physicians’ behavior is on the vertical axis. The three possible physician behaviors are on the horizontal axis. The two curves express the patients’ wishes. Finally, each panel corresponds to one cluster of participants. The effect of the level of social support during hospitalization, the effect of the severity of the disease and the effect of the patient’s psychological robustness were not shown because their impact was generally weak. The distribution of participants in each of the five clusters is shown in Table 1. Separate analyses of variance (ANOVAs) were conducted on the data of each cluster, using a Severity × Wishes × Support × Psychological robustness × Physician’s decision, 3 × 2 × 2 × 2 × 3 design. The detailed results of these ANOVAs can be obtained from the corresponding author.

The first cluster (N = 59, 13% of the sample), was the cluster of people for whom telling the full truth to the patient was the only appropriate decision (M = 9.01). It was called Always Tell the Full Truth to the Patient. Hiding the truth from the patient (M = 2.05), or not telling the full truth to anybody (M = 0.66) was considered as clearly inappropriate, irrespective of the other circumstances (see Figure 1, left panel). As shown in Table 1, French patients (29%) were much more likely to belong to this cluster than Togolese patients (1%).

The second cluster (N = 76, 17%) was called Tell the Truth to the Patient or Relatives because the only appropriate decision for these participants was to tell the truth to the patient (M = 8.38) or her relatives (M = 5.69). This position was, therefore, more nuanced than the previous cluster’s one: Ratings were still high when the physicians tell the truth to the patient’s relatives only. French patients (23%) were more likely to belong to this cluster than Togolese patients (13%).

The third cluster (N = 66, 15%) was called Depends on Patient’s Wishes because the patient’s wishes were the factor that has a major impact on appropriateness judgments. When the patient wished to know the full truth, telling it was considered more appropriate (M = 6.65) than the other options (M = 5.30 and 3.70, respectively). When the patient did not insist on knowing the full truth, not telling it was considered more appropriate (M = 5.77) than the other options (M = 5.08 and 3.30, respectively). Overall, however, telling the full truth was considered as more appropriate (M = 5.98) than the other two options (M = 5.60 and 4.47, respectively). French participants (25%) were more likely to belong to this cluster than Togolese participants (7%).

The fourth cluster (N = 73, 16% of the sample) was a cluster of people for whom hiding the truth from the patient but telling the full truth to her relatives was, in most cases, considered as the most appropriate option (M = 8.71). Telling the full truth directly to the patient was considered as less appropriate (M = 2.0), although hiding the truth was considered totally inappropriate (M = 2.26). This cluster was called Tell the Truth to the Relatives. Togolese participants (27%) were more likely to belong to this cluster than French participants (2%).

Figure 1

In each panel, (a) the judged appropriateness of the physician’s behavior is on the y-axis; (b) the three levels of the physician’s behavior are on the x-axis (DTT = Did not tell the truth to anybody, TTR = Told the truth to the relatives, and TTP = Told the truth to the patient); and (c) the two curves correspond to the two levels of the patient’s wishes.

As expected, putting together and reanalyzing the data gathered by Igier et al. (2015) and by Kpanake et al. (2016) allowed a more fine-grained analysis of people’s positions regarding the breaking of bad news to patients. Six clusters were found that corresponded to five interpretable positions and, for a small group of participants, to the expression of indetermination, whereas only four clusters had been identified by Igier et al. (2015) and only five clusters by Kpanake et al. (2016). Overall, (a) the French participants, more often than the Togolese participants endorsed the view that physicians should tell the truth directly to the patient, particularly when he or she wishes to receive this information, and (b) the Togolese participants, more often than the French participants, endorsed the view that physicians should inform the patient’s family first.

These findings were consistent with previous empirical studies in sub-Saharan African countries that suggested that respect for individual autonomy in health care is not a strongly endorsed value among African people (Kpanake, Dassa, Sorum, & Mullet, 2014; Solum, Maluwa, & Severinsson, 2012; see also Kpanake, 2018). The salience of individual autonomy and self-determination in Western Europe cultural context may explain the French sensitivity to patients’ autonomy in health communication. In contrast, the Togolese preference for family involvement in the breaking of bad news reflects cultural values promoted in Africa, such as interdependence and communalities. These African cultural values may have produced sociocultural norms that do not promote patients’ autonomy regarding the breaking of bad news to patients.

Furthermore, in many African cultures, the break of bad news such as grave illness should be performed by a culturally designated appropriate person, appropriately, at an appropriate time and place (Beyene, 1992). In the Ethiopian cultural context, for instance, "the disclosure of bad news is staged, and close friends and family members prepare persons to face disturbing news by leading them to it gradually. The situation is discussed among friends and relatives to decide the appropriate time and the least frightening way of breaking the news" (Beyene, 1992, p. 330). An additional explanation is that whereas in medically developed countries such as France disclosure of bad news may guide patients throughout numerous treatment options, in the Togolese context such disclosure may undermine patients’ hope since they are aware of the lack therapeutic options.