If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
We analyzed the effectiveness of cross-cultural competence interventions in oncology.
•
In sum, nurses constituted the largest occupational group among participants.
•
The overall effect of interventions was not statistically significant.
•
The dimensions of knowledge and behavior did improve significantly.
•
Effects beyond that remain unclear.
Abstract
Purpose
Cross-cultural competence of healthcare providers is crucial to create a culturally safe environment. Cancer poses special challenges to cross-culturally competent communication and decision-making. Yet, no research synthesis on cross-cultural competence interventions has focused specifically on oncology.
Methods
We conducted a meta-analysis and qualitative review of literature on the effectiveness of cross-cultural competence interventions in oncology. No limitations were placed on publication date, language, oncology setting, or geographic region. Of 1.565 citations identified, 15 articles met the inclusion criteria. Information on study design, samples, measured outcomes, and effectiveness statistics were coded. Average weighted effects were calculated applying meta-analysis methodology.
Results
Studies were published between 2000 and 2020; more than half in the last seven years; two thirds in the USA. Overall study quality was at a low to moderate level, notably only one study provided a control-group-design. In sum, nurses constituted the largest occupational group among participants. Results of the meta-analysis indicate that cross-cultural competence interventions have differential effects. While the overall effect of cross-cultural competence interventions was not statistically significant, results showed that the cross-cultural competence dimensions of knowledge and behavior did improve. Effects beyond that remain unclear.
Conclusions
We provide valuable information on research gaps. The lack of studies and insufficient methodological rigor of available studies show that more research is needed to support the claim that interventions actually improve the various dimensions of cross-cultural competence in oncology. To build a stronger evidence base, it is necessary to include patient-reported outcomes and to center their experiences in future research.
In 2019, the number of migrants worldwide reached 292 million, with Europe ranking first among destination regions, followed by Northern America; in addition, the number of migrants (non European Union (EU) migrants and mobile EU citizens) living in the EU-27 Member States (including UK) has increased by around 60% over the last 20 years (
). While migration itself is not a risk factor for health problems, the conditions that accompany migration processes often increase the risk for poor health (
International Organization for Migration Integrating Migration Into Health Interventions: A Toolkit For International Cooperation And Development Actors, UN Migration.
). Studies indicate that some migrant groups such as children and elderly persons, exhibit higher rates of health problems (e.g., higher hospitalization rate, higher obesity rate) compared to the non-migrant population (
). These problems may be attributed to the following factors: poorer healthcare access in the home country, lower level of education, traumatic experiences during the migration process, poor health literacy and limited knowledge of the structure of the health system in the host country, poor communication between healthcare professionals and migrant population (
International Organization for Migration Integrating Migration Into Health Interventions: A Toolkit For International Cooperation And Development Actors, UN Migration.
International Organization for Migration Integrating Migration Into Health Interventions: A Toolkit For International Cooperation And Development Actors, UN Migration.
)) to improve access to and quality of healthcare for patients with a migrant background.
According to the World Health Organization (WHO), cancer is the most common cause of death worldwide. More specifically, one in six deaths in 2020 were due to cancer (
). This positive development, brings new challenges both to patients (e.g., balancing ongoing obligations such as work and inpatient cancer treatment and uncertain future, psychological burden, sustaining normality) (
) and healthcare professionals (e.g., coordination of follow-up cancer care and support services). Many of these challenges are further exacerbated in the case of cancer patients with a migrant background due to, for example, patients' limited familiarity with the healthcare system or barriers to patient-provider communication (
International Organization for Migration Integrating Migration Into Health Interventions: A Toolkit For International Cooperation And Development Actors, UN Migration.
). Studies in fact indicate that oncology care providers often face difficulties due to language barriers and cultural and religious differences (e.g., families’ religiously and culturally determined understandings of illness) (
Oncology care needs to cater to each patient's cultural beliefs and values. As the proportion of culturally diverse patients in oncology is likely to increase (
) to create a culturally safe environment and enable shared decision-making (SDM) between the involved parties. SDM is a collaborative process in which healthcare professionals work together with patients to reach a consensual decision about care based on treatment options, outcomes, and patient background, needs, and preferences (
). At a conceptual level, cross-cultural competence (CCC) is a broad term that encompasses various theories and models aimed at better understanding one's own culture, developing positive attitudes towards other cultures, and improving interaction with them (
). In addition, CCC facilitates culturally congruent care; a process of matching families' needs and preferences with healthcare professionals' knowledge, attitudes, and skills, thereby recognizing that care actualizes as a result of dynamic interactions between patients and healthcare professionals (
). CCC is widely recognized as being essential to reduce health disparities, to deliver equitable care to all patients independently of their cultural or religious background, and to ensure cultural safety in clinical practice (
). Beach et al. highlight that CCC interventions show potential in improving health professionals' knowledge, attitudes and skills, as well as patient satisfaction (
). In addition, Truong et al. point out that although there is some evidence that CCC interventions improve patient health outcomes, the included studies often have methodological flaws and that many of the studies rely on self-reports (i.e., by providers) on the effectiveness of the intervention, which increases the risk of biases (
). Despite the fact that cancer poses special challenges (e.g., end-of-life decision-making, describing multifactorial etiologies, often manageable chronic disease, palliative care) to cross-culturally competent communication and decision-making (
). Hence, we conducted a meta-analysis and qualitative review of the literature on interventions designed to improve CCC in oncology to determine their effectiveness.
2. Methods
2.1 Search methodology
To frame the review, Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were followed (
). Using Boolean logic (Table 1), we searched the following six databases: Scopus, PubMed, PsycInfo, Web of Science, Cinahl, and SocIndex. The search resulted in a total of 1.565 studies (Table 1). Moreover, we contacted authors to obtain unreported data, screened references of included articles for relevant additions, and attempted to solicit additional citations through discussion with topic experts.
Table 1Search terms and search results.
Search terms
Matches
Scopus
PubMed
Psyc- Info
Web of Science
Cinahl
Soc Index
1)
“Cultural competenc*” OR “Cross-cultural competenc*” OR “Intercultural competenc*” OR “Transcultural competenc*” OR “multicultural competenc*”
15.327
8.563
8.188
9.627
11.735
2.766
2)
Cancer OR oncolog*
3.446.536
4.543.118
68.965
2.807.808
500.318
19.128
3)
Indicator* OR measure* OR intervention* OR program* OR evaluation* OR training* OR teaching OR tutoring OR assessment* OR strategy OR strategies OR effective* OR efficacy OR prevent* OR instruction* OR coaching
We defined the following inclusion criteria: (a) peer-reviewed research articles had to (b) report quantitative evaluations of (c) interventions that (d) aimed to enhance CCC of (e) oncology providers. No limitations were placed on publication date, language, oncology setting (adult, pediatric), and geographical region.
2.3 Search results, data extraction, and data coding
After de-duplication, 821 articles remained. Two researchers (first and last author) independently reviewed titles and abstracts, removing those not meeting the inclusion criteria. This step resulted in 35 articles selected for full-text review, in addition to four articles identified from cited references; 15 remained after full-text review and comprised the final set of included studies (Fig. 1, Table 2). Records were managed through the software EndNote.
We developed a 80-item codebook containing information surrounding the following main areas of interest: description of publication (e.g., year, country, journal), study design (e.g., control group, pre-post design, time between measurement points), sample characteristics (e.g., sample size, attrition, occupational groups, work-experience, age), intervention (e.g., objective, underlying theoretical CCC framework, voluntariness, length, group size), outcomes (e.g., measures, measurement method, standardized tool), and statistics (e.g., mean and standard deviation of pre- and post-tests, t-value, F-value, p-value, effect sizes). Given the small number of included studies, the first and last author coded the set of studies together. Thus, calculation of interrater-agreement was not needed. Data were managed through SPSS.26. In addition to coded quantitative data, available qualitative data was extracted using a purpose-build Excel extraction form, which captured the following areas of interest: providers’ perceptions of benefits and challenges of the intervention, overall satisfaction, concerns, transferability (e.g., other clinical contexts, other minority populations), and recommendations.
2.4 Computation of effect sizes and statistical analyses
Since none of the included studies provided a controlled evaluation design, effect sizes had to be calculated based on within-person repeated measures. In this case, a standardized mean change is calculated from pre and post means and standard deviations (
). Furthermore, the correlation between the pre and post measurement should be taken into account for the effect size estimation. If this correlation is not available, the authors recommend setting it to 0. If primary statistics (mean, standard deviation, and potentially correlation) were not reported, transformation procedures (
) were applied to the given test statistics (e.g., t-value, p-value). The computed effect sizes can be interpreted in a similar manner as Cohen's d, thus an effect size of 0.2 is considered small, 0.5 as medium, and 0.8 as large (
). Random effect models were fitted by applying restricted maximum likelihood estimation, and Forest plots were constructed to visualize the results. Statistical procedures were executed with the metafor package in R-statistical software (
). Based on the method of moments point estimation, random effect models were applied to estimate weighted mean effect sizes. A two-sided P-value ≤0.05 was considered significant for all analyses.
2.5 Quality appraisal
We assessed study quality using the Design and Implementation Assessment Device (DIAD), which is specifically tailored for intervention effectiveness research (
A systematic and transparent approach for assessing the methodological quality of intervention effectiveness research: the Study Design and Implementation Assessment Device (Study DIAD).
). DIAD does not provide a single number to represent the multidimensional construct of methodological study quality, but offers a profile of quality scores, namely fit between concepts and operations, clarity of causal inference, generality of findings, precision of outcome estimation (Table 3, appendix). The quality of one study could not be assessed, since assessment tool and study design did not match.
3. Results
3.1 Characteristics of included studies and study quality
All studies were published between 2000 and 2020. Two thirds of the studies were carried out in the USA; it is noteworthy that the first non-US study was only published in 2016, the first (and only) European only in 2020. More than half of the studies (i.e., 8/15) were published in the last seven years. Mean sample size was 63.8 providers (standard deviation = 52.4). Ten studies were publicly funded. All studies were conducted in the adult oncology setting. With respect to occupational groups within a sample, four were exclusively comprised of nurses, who were the largest occupational group in four more studies. Eight studies referred to the underlying conceptual framework of CCC. In nine studies, the evaluators had developed the intervention themselves. Eight studies indicated the overall duration of the intervention (mean = 8.3 h, standard deviation = 10.5 h). Only one study assessed effectiveness of the intervention from a patient perspective. Finally, there was no variation in World Bank income classification as all studies were from high income countries.
Many of the 35 individual items of DIAD could not be answered with certainty due to non-reported information. This resulted in a high number of “maybe no” answers to the composite questions of the four DIAD scores. Overall, due to non-reported information and severe methodological limitations, the quality of included studies was low to moderate. As such, fit between concepts and operations, clarity of causal inference, generality of findings, and precision of outcome estimation could mostly not be confirmed. Main methodological limitations included the following: only one study used an untreated control group, three studies did not employ pre- and post-tests of providers’ CCC, only two studies had a follow-up assessment, four studies did not report attrition rate, in all studies voluntary participation possibly biased effectiveness estimations (e.g., ceiling effect), only two studies described the intervention at a level of detail that would allow its replication, and only six studies employed standardized tools for which quality criteria were known to measure their outcomes.
3.2 Mean effects
Only nine of the 15 included studies provided sufficient statistical information so that an effect size could be obtained. The average weighted effects are displayed in Fig. 2. The overall effect has to be considered a null effect since the effect size was not significantly different from 0. In the sub-categories, however, significant average effects could be found with regard to knowledge and behavior outcomes (Table 4).
Notes. Average weighted effects were calculated according to the random effect size model. k = number of studies providing an effect size within this outcome category.
To explore bias due to the so-called file-drawer problem (less conclusive studies or studies with small effect are less likely to be published) a linear regression from sample size to the unweighted effect sizes was computed (
). A non-significant slope indicates the absences of a publication bias ((β = 0.00, t = −0.195, p = 0.851). Begg's rank correlation was non-significant as well (rtau = −0.17, p = 0.612, (
). A funnel plot (Fig. 3, appendix) with a subsequent trim and fill analysis (Fig. 4, appendix) yielded the estimation of only three missing studies on the left side (
). The corrected mean effect size was then estimated to be 0.199 (p = 0.231) and again non-significant. This finding was further supported by a significant regression test for funnel plot asymmetry (t = 29.62, df = 7, p < 0.001, (
)) indicating a slight bias due to three missing studies.
4.1 Additional qualitative results
Qualitative results indicate that the participants value courses on CCC, because there is an exchange with other participants and the methods learned in the courses might be transferable to other population groups. Results further show that participants have ongoing training needs (e.g., supervision; hands-on), and that a course should be adapted to participants' level of pre-existing CCC. Furthermore, participants wished for patient feedback and debriefings with other team members (Table 5). Practicing cultural humility was a common concern among course participants.
Training community practitioners in a research intervention: practice examples at the intersection of cancer, Western science, and native Hawaiian healing.
Although participants reported increased knowledge, many of them perceived a certain discomfort in delivering the ho'oponopono procedures as they were concerned of not being credible leaders. They felt either not mature or not knowledgeable enough. Some expressed also concern about not being Hawaiian themselves. More practice and supervision were seen as possible solutions to these problems.
Participants stated that they gained an increase in knowledge or skills through the training. In addition, participants acknowledged that some of the content could be applied to other cultures (e.g., information on sensitivity and respect).
Participants appreciated the hands-on experience through the Observed Standardized Clinical Experience and the exchange with course instructors and other participants. Moreover, participants reported that they did not feel alone and found it stimulating to learn from others facing similar kinds of difficulties.
Evaluation from patient perspective: Indigenous cancer patients are overall very positive about their experience with the indigenous patient navigator. They appreciated the fact that they could talk with someone and that the indigenous patient navigator showed compassion.
Nurses were satisfied with the Indigenous Relationship and Cultural Safety courses in terms of knowledge transfer, content delivery, delivery method (i.e., online and self-pace). Although they believed the courses to be especially beneficial for nurses with limited work-experience with First Nations, Inuit, and Métis people, also those with a lot of experience with First Nation communities or were indigenous themselves, found the courses valuable especially regarding cultural safety.
Some nurses highlighted that although the courses are a good beginning, they are not enough to foster cultural humility (rather than competence/knowledge).
Providers were divided into low and high cultural competence (CC) groups. Providers in the low and high CC group showed important differences regarding content of previous CC training, impact of previous cultural competency training on behavioral changes during interactions with patients and colleagues and ongoing training needs and structural support.
Overall providers in the high CC group had received mostly trainings that were scenario-based, whereas the training materials of those in the low CC group were more definition-based and culture-specific with little or no focus on real practice interaction with patients. Providers in the high CC were more open to self-reflection and self-critique and thus more focused on self-improvement by engaging in dialogues with colleagues and patients.
Those in the low CC group tended to focus on body language and altered their communication methods but were less cognizant of their own biases and tended to attribute patient behavior to culture rather than to situational factors. Although all providers desired ongoing training, those in the low CC group wished for more standard cultural decision-making tools, while those in the high CC group desired more patient feedback and debriefings with other team members.
Developing strategies for reducing cancer disparities via cross-institutional collaboration: outreach efforts for the partnership between the Ponce School of Medicine and the Moffitt Cancer Center.
Communication about breast cancer genetic counseling with patients with limited health literacy or a migrant background: evaluation of a training program for healthcare professionals.
To our knowledge, this is the first study that provides a systematic synthesis of the effectiveness of CCC interventions in the oncology setting. Results indicate that CCC interventions have differential effects. While the overall effect of CCC interventions was not statistically significant, results showed that participants' knowledge and behavior could be improved. Even though effect sizes have to be considered as rather small and the number of included studies was low, it is notable that behavior changes generated the highest effects. Our findings (e.g., low study quality, low number of studies, only high-income countries, no causal inference possible), however, cannot confirm the widely held belief that CCC interventions effectively enhance (all dimensions of) oncology providers’ CCC. Most studies were published only after 2016, which points towards a fairly new interest in the subject, possibly due to demographic changes and raised awareness for cultural safety in healthcare, especially outside of the USA where the first study was published only in 2016.
Our meta-analysis provides valuable information on where future research in the field needs to fill the gaps. Generally, more research is needed to draw robust conclusions on the effectiveness of interventions to enhance CCC in oncology. Furthermore, studies applying a randomized controlled design are lacking entirely. The latter needs to be addressed with the highest priority, since only under these controlled conditions causal inference is truly possible (
). Future studies need to measure attitudes and awareness, since, for these dimensions, conclusions from the present meta-analysis are very weak owing to the small number of studies. Due to the unique triadic decision-making constellation in pediatrics (i.e., child, family, providers), studies need to be carried out in the pediatric setting as well. Future research would greatly benefit from a uniform conceptual model of CCC and from using standardized, validated tools to measure CCC; these shortcomings have been lamented for years (
). Moreover, since curricula and training methods as well as sample compositions (i.e., proportions of different occupational groups) vary across interventions, future research syntheses should account for these differences. Lastly, the majority of studies stem from the US oncology setting and all studies were carried out in high income countries. Hence, non-US research is urgently needed.
For the oncology setting, recent findings suggest that there are differences in CCC among occupational groups and that these differences vary across dimensions of CCC (
). Our results of differential effects across dimensions of CCC, thus, might be explained by different compositions of the respective study samples (which we could not address within a moderator analysis owing to a low number of studies), as suggested by previous research (
). In conjunction, these findings corroborate the need to, first, determine occupational groups' different levels of CCC dimensions and to, second, specifically tailor interventions to the respective occupational group's level of CCC. Otherwise, financial and time resources might be wasted and some providers might experience cognitive underload and a lack of motivation, since interventions do not offer much new information. Similarly, the qualitative results of the included studies also described the need to adapt the interventions to the already existing CCC levels of the providers.
Our analysis yields no statistically significant increases in overall CCC. Increases in knowledge and behavior are partly congruent with other studies’ results. While there is robust evidence that interventions improve knowledge and behavior (
), previous findings on overall CCC showed no increase in treatment-control design studies, but an increase in pretest-posttest effect size synthesis (
), which is confirmed by our results. Furthermore, contrary to our findings, previous research revealed beneficial effects of interventions on attitudes (
). Given the low number of included studies, particularly for attitudes, results of our meta-analysis should be interpreted with strong caution. Finally, in line with other research, qualitative findings indicate that providers generally appreciated CCC interventions (
Culturally competent healthcare - a scoping review of strategies implemented in healthcare organizations and a model of culturally competent healthcare provision.
), in particular the exchange with other participants and transferability of learnings. In addition, providers emphasized that the acquisition of CCC is an ongoing process that requires further courses and that cultural humility requires more than (interventions on) CCC. The latter mirrors a line of conceptual critique of cultural competence (
First, the lack of randomized controlled trials studies is a threat to the causal interpretations from the meta-analytic results. Cook and Campbell even argued that such results are uninterpretable (
). Yet, in this particular case, it is not very likely that the effects are due to maturation of the participants. Thus, it can be assumed that the significant average effects stem, at least to some extent, from the CCC intervention. Second, due to the low number of included studies, we could not perform moderator analysis. Previous research has shown that effect sizes vary as a function of numerous variables (e.g., funding source, provider types) (
). Third, we only included published results, raising the issue of publication bias. Fourth, since only one study reported patient-related outcomes, our findings do not reflect effectiveness of interventions as perceived by patients.
6. Conclusions
Findings suggest positive effects of interventions on cross-cultural knowledge and behavior. The effects beyond that remain unclear. Given the lack of studies examining the effectiveness of CCC interventions in oncology and the insufficient methodological rigor of the available studies, more thorough research is urgently needed to support the claim that interventions actually improve CCC in oncology and to better understand which cross-cultural dimensions can be most effectively improved. However, building a stronger evidence base requires not losing sight of patient-reported outcomes and putting their experiences at the centre of future research.
Funding
This work was supported by the Swiss Cancer League (grant number: KLS-4822-08-2019).
Declaration of competing interest
All authors declare that they have no competing interests.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
Developing strategies for reducing cancer disparities via cross-institutional collaboration: outreach efforts for the partnership between the Ponce School of Medicine and the Moffitt Cancer Center.
Culturally competent healthcare - a scoping review of strategies implemented in healthcare organizations and a model of culturally competent healthcare provision.
Training community practitioners in a research intervention: practice examples at the intersection of cancer, Western science, and native Hawaiian healing.
A systematic and transparent approach for assessing the methodological quality of intervention effectiveness research: the Study Design and Implementation Assessment Device (Study DIAD).
Communication about breast cancer genetic counseling with patients with limited health literacy or a migrant background: evaluation of a training program for healthcare professionals.
00011-X&id=gr1.jpg){kind=link}
00011-X&id=gr2.jpg){kind=link}