Assessing delirium knowledge among health care professionals: findings from a scoping review

The aims of this study were to map: (1) studies assessing delirium knowledge among HPs; (2) instruments, questionnaires, types of item or methods used to date (hereinafter, instruments), their validity and reliability; and (3) domains regarding delirium prevention and management knowledge mostly investigated among HPs to date.

Our study has been guided by the following research questions: (1) What studies focused on delirium knowledge among HPs have been published to date? (2) What are the main characteristics of the instruments used to date to assess delirium knowledge among HPs? (3) Which domains of delirium knowledge among HPs have been the most investigated to date?

This scoping review was guided by the methodological framework of Arksey and O’Malley [22] and Levac et al. [23]. After identifying the research questions, the following four stages were identified: (a) identifying relevant studies; (b) selecting studies; (c) charting data; and (d) collating, summarizing, and reporting the study findings. The systematic review process was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR; [24], as summarized in Additional File 1.

The search strategies reported in Additional File 2 were applied to Medline, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Embase, PsycINFO, and Scopus. Eligible studies were those that satisfied the following inclusion criteria: a) involving HPs; b) exploring knowledge among HPs of delirium in adult and/or paediatric patients; c) in all settings (hospital, primary, and/or long-term care settings); d) use of any type of instrument or method to assess the knowledge; e) validating instruments to assess the knowledge possessed by HPs regarding delirium; and f) at any time.

We excluded studies: a) assessing delirium prevalence, incidence, aetiology, and/or treatment(s); b) case reports, editorials or letters to the editor, conference abstracts, or posters; and c) not reporting the instruments in the publication.

Two researchers (CM and IM) conducted the literature search and worked independently to evaluate studies’ eligibility based on screening of the titles and abstracts retrieved. Differences were discussed with a third researcher (DA). The full texts of the eligible studies were then retrieved. By reading the full texts, two researchers (CM and IM) independently assessed their eligibility; any decision on article inclusion was then based on joint agreement. Figure 1 shows the process of study selection and inclusion, following the PRISMA-ScR guidelines.

Two researchers (CM and IM) piloted the extraction grid with ten studies and no changes were necessary. The following data were extracted from each included study: first author; year of publication; continent of author’s affiliation; aims and study design; participants involved and main characteristics; instruments used to assess delirium knowledge by assessing if developed/adapted from a previous instrument; number and type of items (e.g. questions, vignettes); type of answers (e.g. dichotomous, multiple choice, Likert scale, open answer); and psychometric properties resulting from validation processes, if available.

Secondly, the available instruments were examined (CM, IM), extracting all items, which were reported in a grid designed to collect all items retrieved. This grid was piloted with ten studies and no changes were needed. The researchers worked independently and compared the extracted data. All differences were discussed with a senior researcher (AP).

The following data analysis was performed.

A total of 98 studies were included (Table 1). The majority were pre-post (n = 47, 48.0%) (e.g. [28] and cross-sectional studies (n = 42, 42.9%) (e.g. [29]), mainly involving nurses (n = 57, 58.8%). The majority were performed in America/Canada (n = 34, 34.7%; e.g. [30]), Europe (n = 29, 29.6%; [31]) and Asia (n = 25, 25.5%; e.g. [32]). The sample sizes ranged from ten [33] to 1,215 participants [31], mainly using convenience sampling (e.g. [34, 35]).

Several studies (n = 44, 45.4%) were intended to assess knowledge, attitudes, and abilities regarding delirium prevention and management (e.g. [36]). Others (n = 42, 43.3%) aimed at evaluating the effectiveness of specific educational interventions on delirium knowledge (e.g. [37]), assessing the perceived barriers associated with delirium detection/screening tool use (n = 7, 7.2%) (e.g. [38]), or testing the validity and reliability of instruments designed to assess delirium knowledge (n = 4, 4.1%) (e.g. [39]).

However, most studies (n = 54, 55.2%) (e.g. [40]) did not report the healthcare setting, whereas 33% (n = 32) were performed in critical care units (e.g. [41]). A few studies were performed in non-critical hospital units (n = 8, 8.2%) (e.g. [42]), nursing homes, or other long-term care settings (n = 4, 4.1%) (e.g. [43]).

Of the 98 included studies, at least one instrument was used in 43, and a total of 32 different instruments, as summarized in Table 2. The majority were self-developed by authors (n = 17, 53.1%) (e.g. [44]), whereas among the others the most used (both in their original form and with modifications) were the Delirium Knowledge Questionnaire developed by Hare et al. [45] (14 studies; e.g. [46]); the 35-item true–false Delirium Knowledge Questionnaire from Detroyer et al. [47] (four studies; e.g. [13]); and the questionnaire from Davis and MacLullich [48] (three studies; e.g. [31]). The vignettes developed by Fick et al. [49] were used in three studies (e.g. [50]).

Multiple types of item were used in 98 studies. Of the 32 available instruments, 78.1% included questions (n = 25), whereas fewer used vignettes (n = 3, 9.4%) and clinical cases (n = 2, 6.3) (Additional File 3). Some instruments included multiple types of item (n = 2, 6.3%), such as a combination of questions and clinical cases or vignettes (e.g. [51]).

In the case of instruments with items, the average number of items is 22 (interquartile range 12.0–32.2). Different modalities of answers are used (e.g. multiple choice and Likert scales) (e.g. [35]).

Evidence of validity was reported for a few instruments (11 out of 32; Additional File 4). The instruments’ reliability (e.g. internal consistency) and validity (e.g. content validity, construct validity) were reported in only one study [52], following the COSMIN taxonomy [53]. In six studies (18.8%) (e.g. [34]), only some psychometric properties were reported (e.g. internal consistency with the Cronbach’s alpha), while four (15.6%) (e.g. [35]) referred to previous validity and reliability tests (data are summarized in Additional File 4) [54‐59].

As summarized in Table 3, all items extracted from the instruments (n = 392) were categorized into domains and sub-domains. In the first domain, ‘Definition, signs and symptoms’, 81 items were categorized (20.6%); in the second, ‘Risk factors, incidence and prevention’, there were 139 items (35.4%); in the third, ‘Detection and tools’, there were 89 items (22.7%); in the fourth, ‘Management and therapy’, there were 64 items (16.3%); and in the fifth, ‘Outcome, prognosis and consequences’, 18 items were categorized (4.8%).

A total of 98 studies have been published to date, the first published in 2008 [45], indicating that this research field has been established in the last 15 years. Moreover, in terms of production, an average of 6.5 articles a year has been produced, mainly in specific continents (America, Canada, and Asia). In addition, approximately half of the studies were performed to measure the effectiveness of educational strategies on knowledge with pre-/post- designs, and the other half investigated HPs’ knowledge with cross-sectional designs. Therefore, this field of research has been moderately productive in the last 15 years, specifically in some countries, with a prevailing descriptive intent.

Although patient safety, including the prevention and management of delirium, is the responsibility of all HPs [1], almost 60% of the studies investigated nurses’ knowledge only [18, 20]. Nurses are providing direct care and are well positioned to assess and monitor patients’ conditions, and thus to identify potential changes in mental status associated with a delirium. The early detection of delirium by using screening tools is an effective strategy to decrease adverse events [61]. Therefore, nurses need to possess the required knowledge to prevent and detect early delirium episodes. In recent years, delirium has been integrated into the curriculum for undergraduate physicians, recognizing the importance of delirium as a “core business” of future physicians [62]. As the NICE guidelines recommend delirium prevention and management are the responsibility of multidisciplinary teams and require an interprofessional approach [1], where clinical decisions are shared among different HPs (e.g. physicians, nurses, physiotherapists). Outcomes can be improved when patients with delirium are cared for by appropriately trained interdisciplinary teams [62]. However, there is still little evidence on the effectiveness of interprofessional delirium programmes in daily practice, including educational interventions [63]. Greater efforts in research and practice should be made to create awareness and to assess delirium knowledge among all HPs involved in patient care, as an important component in preventing and managing this clinical complication [61].

One-third of the studies involved HPs employed in intensive care units, while only four regarded long-term care settings, including nursing homes. Moreover, several studies did not report on the setting in which they were conducted. Delirium is a common complication in acute care, including intensive and post-operative care. However, in long-term care settings, patients often develop delirium for a longer period and have worse outcomes [64, 65]. Long-term care settings are still neglected [66, 67], suggesting a priority in future research, given the risk of underestimating episodes of delirium according to the complexity of the identification associated with the occurrence of dementia [49].

A total of 32 different instruments emerged, suggesting that this field of research is very rich in terms of the tools produced. Among the instruments most used, the Delirium Knowledge Questionnaire [45] was most popular.

From a conceptual point of view, by considering the Dublin descriptors (European Higher Education Area, 1998), instruments used to date seem mainly to assess knowledge and understanding using items (e.g. [45]); in a few studies, the capacity to apply the knowledge in given cases (e.g. vignettes and clinical situations; [51]) has been investigated. Issues in HPs’ knowledge have been highlighted as a possible reason for the lack of preventive and managerial interventions for delirium [12]. However, while academic or theoretical knowledge can be important in the context of HPs (e.g. at the undergraduate level), practical knowledge has more relevance given its necessary precondition for transitioning knowledge into practice. Therefore, alongside a consensus regarding the best tool to be used in the future according to the more complex learning and teaching taxonomy outcomes [68], increased attention is recommended towards tools able to measure the applied knowledge, analyse the situation, and help critically evaluate and synthesize results. Indeed, according to the knowledge-attitude-practice theory, simple knowledge acquired is not able to address changes in behaviour [69].

No psychometric properties have been documented for 21 of the instruments, and around half of them have been self-developed by authors. In the remaining 11, the validity and reliability of data obtained in this research area seem to be limited, requiring further investigation. The most investigated psychometric properties are face validity and internal consistency. Only the study [52] reported more detailed evidence based on validity and reliability to assess delirium knowledge among critical care nurses; in the other ten studies, different validity measures were assessed with ten different tools, suggesting fragmented evidence. Specifically, regarding validity, five studies out of 11 (e.g. [47]) have assessed face validity, suggesting more effort in evaluating the capacity of the instrument to assess what it is intended to evaluate [70]. In addition, four authors reported content validity data expressed with the content validity index [39, 52, 57, 71], considering the degree to which a sample of items, taken together, constitute an adequate operational definition of a construct [72]. In all studies, the index was higher than the recommended threshold of 0.080 [73]. Regarding reliability, five studies assessed Cronbach’s alpha [34, 35, 41, 44, 57], four reporting an acceptable value ≥ 0.70 [74]. Consequently, a more rigorous approach is required to produce valid measures in the field with a more cumulative approach by assessing the properties of the same instrument to increase its validity progressively.

The findings fall into two main categories.

The different instruments are focused on some domains and sub-domains and neglect the importance of others. Given that no instrument can be considered as the gold standard, the emerging map may in future address the development of more comprehensive tools assessing knowledge of all-important aspects of delirium also according to the different levels of education [78, 79].

The scoping review has several limitations. Firstly, no grey literature or unpublished studies have been considered, and this may have introduced selection bias. Secondly, despite experts in the field of delirium and research methodology having been involved, and data extraction and analysis having strictly followed the research questions, their categorization might have been influenced by the background of the researchers. Thirdly, the process of categorization of the items extracted from the instruments has been performed by combining two approaches, both according to a given classification [25, 26] and inductively (item by item analysis). The map that emerged may be influenced by the process used; moreover, the vignettes (see Additional File 3) have not been categorized given their different features.