Study characteristics, research methods, and interventions varied among the 18 studies. In this section, study characteristics, including locale, setting and sample, intervention structure and approach, conceptual underpinning, measures, study designs, and study global quality ratings, are summarized and synthesized.
Locale. Of the 18 studies, 11 studies took place in the United States (Beer, Hutchinson, & Skala-Cordes, 2012; Burgio et al., 2001; de Abreu, Hinojosa-Lindsey, & Asghar-Ali, 2017; Fritsch et al., 2009; George, Stuckey, & Whitehead, 2014; Levy-Storms, Harris, & Chen, 2016; McCallion, Toseland, Lacey, & Banks, 1999; Passalacqua & Harwood, 2012; Ripich, Wykle, & Niles, 1995; Teri, Huda, Gibbons, Young, & Leynseele, 2005; Williams, Perkhounkova, Herman, & Bossen, 2017). Seven studies took place in other countries, including the United Kingdom (Adefila, Graham, Clouder, Bluteau, & Ball, 2016; Cockbain, Thompson, Salisbury, Mitter, & Martos, 2015), The Netherlands (Sprangers, Dijkstra, & Romijn-Luijten, 2015), Australia (Broughton et al., 2011; Conway & Chenery, 2016), and Canada (Eritz et al., 2016). Hattink et al. (2015) included participants from the United Kingdom and The Netherlands.
Setting and Sample. Ten studies took place in or recruited staff from LTC facilities (Broughton et al., 2011; Burgio et al., 2001; Eritz et al., 2016; Fritsch et al., 2009; Levy-Storms et al., 2016; McCallion et al., 1999; Passalacqua & Harwood, 2012; Ripich et al., 1995; Sprangers et al., 2015; Williams et al., 2017). Two studies involved home-based care or community-dwelling individuals (Conway & Chenery, 2016; Hattink et al., 2015) and only one involved assisted living residents (Teri et al., 2005). Five studies enrolled students at academic settings or universities (Adefila et al., 2016; Beer et al., 2012; Cockbain et al., 2015; de Abreu et al., 2017; George et al., 2014). Two studies assessed the same intervention but in different health care contexts— nursing homes and community aged care centers, respectively (Broughton et al., 2011; Conway & Chenery, 2016).
Care provider sample sizes ranged from 16 (Levy-Storms et al., 2016) to 192 (Fritsch et al., 2009) at post-intervention analysis. Sample size was often even smaller at follow-up assessments. Seven studies specify the inclusion of nursing assistants (NAs) or certified nursing assistants (CNAs) in their study population (Burgio et al., 2001; Fritsch et al., 2009; Levy-Storms et al., 2012; McCallion et al., 1999; Ripich et al., 1995; Sprangers et al., 2015; Williams et al., 2017). Details of the study samples are shown in Table C.
Five studies used student participants, specifically those preparing to enter fields that included dementia care (Adefila et al., 2016; Beer et al., 2012; Cockbain et al., 2015; de Abreu et al., 2017; George et al., 2014). Two studies included students in medical school (de Abreu et al., 2017; George et al., 2014), one included students studying medicine in undergraduate school (Cockbain et al., 2015), one included students receiving higher degrees in a variety of medical fields (Adefila et al., 2016), and one included students in a NA training program (Beer et al., 2012).
Nine studies included actual patients or residents (range = 16 to 105) residing in a LTC facility or assisted living facilities with moderate to severe levels of cognitive impairment. Burgio et al. (2001) included 67 patients with a mean Mini-Mental State Examination (MMSE) score of 13.39; McCallion et al. (1999) included 105 patients with a mean MMSE score of 4.9 in the control group and 6.3 in the treatment group; Sprangers et al. (2015) included 26 residents with a mean MMSE score of 11.17; and Teri et al. (2005) included 25 residents with a mean MMSE score of 16 and problems with depression, anxiety, or agitation. The remaining three studies used a variety of approaches to assess cognitive function, but all three included a mix of residents with moderate and severe levels of cognitive impairment (Eritz et al., 2016; Levy-Storms et al., 2016; Williams et al., 2017). Two studies did not provide details regarding the number of patient participants included or cognitive function status (Fritsch et al., 2019; George et al., 2014).
Intervention Structure and Approach.Table C provides details about the study intervention components including the number of intervention sessions, training approach, and an overview of the content. To summarize, 15 of 18 interventions included multiple sessions, and 13 of 18 studies included an intervention with high investment of time to implement (i.e., >60 minutes). Six studies focused on communication skills training. Some studies taught skills in a lecture format, whereas others used multimedia components, such as DVDs, or interactive components, such as role-play, scripted vignettes, and facilitated group discussions. Hattink et al. (2015) provided training entirely on the internet in a self-directed program in which care providers watched videos, completed interactive exercises, interacted with a learning path advisor (i.e., learning facilitator), and participated in an online social community. Other studies used a variety of approaches, including a simulation of what it is like to live with dementia (Adefila et al., 2016; de Abreu et al., 2017).
The simulation intervention used by de Abreu et al. (2017) attempted to enhance attitudes among U.S.–based psychiatry residents and medical students, specifically their person-centeredness. Adefila et al. (2016) studied virtual reality experiential learning on empathy, defined as compassion, in U.K. health professional students. It is significant that both studies used simulation to promote empathic understanding or imagining of the cognitive and sensory impairments persons with dementia struggle with on a daily basis. In addition, these studies included learning about the person by attempting to see the world from his/her perspective, which is a key component of empathy (Adefila et al., 2016; de Abreau et al., 2017).
Two studies evaluated the impact of a group story-telling program (Fritsch et al., 2009; George et al., 2014). George et al. (2014) also incorporated an educational classroom component to teach the training and a care method known as validation therapy. Finally, Eritz et al. (2016) assessed the use of patient life histories in care provision, in which care providers were given information and participated in interactive discussion with other staff members.
Studies Conceptual Underpinning. Kitwood's (1997) person-centered approach for dementia care was implicitly or explicitly acknowledged in several studies. Kitwood (1997, p. 89) identified 12 “positive person work(s)” to guide care providers involved in dementia care to enhance personhood, with each one enhancing personhood in a different way. Interventions were assessed for inclusion of two of these positive interactions (i.e., recognition and negotiation) (Kitwood, 1997). Eleven of 18 interventions included a component of staff-initiated recognition by specifically making and/or maintaining eye contact (Beer et al., 2012; Broughton et al., 2011; Cockbain et al., 2015; Fritsch et al., 2009; Levy-Storms et al., 2016; Ripich et al., 1995; Teri et al., 2005; Williams et al., 2017), consideration of the care provider's facial expression (Conway & Chenery, 2016; Teri et al., 2005), and showing interest through listening or body mechanics (Beer et al., 2012; Conway et al., 2015; Fritsch et al., 2009; George et al., 2014; Levy-Storms et al., 2016; Passalacqua & Harwood, 2012; Teri et al., 2005). Ten interventions included a component of staff-initiated negotiating (i.e., putting assumptions aside, demonstrated by asking and listening to the care recipient, and learning about preferences, desires, and needs of the person with dementia). Of the 10 studies, two used simulation to learn about the experiences and challenges of a person with cognitive and sensory impairments (Adefila et al., 2016; de Abreu et al., 2017). Eight studies included learning about the background of the resident, including the use of photographs and memorabilia to promote conversations and improve care (Beer et al., 2012; Broughton et al., 2011; Burgio et al., 2001; Conway et al., 2015; Eritz et al., 2016; McCallion et al., 1999; Sprangers et al., 2015; Teri et al., 2005). An examination of the intervention programs focused on Kitwood's (1997) positive interactions revealed that recognition (i.e., gaining and/or maintaining eye contact and other nonverbal communication) and/or negotiating (i.e., asking or learning about preferences, desires, and needs) were techniques used in most studies.
Measures. There were a number of measures used to assess the effectiveness of the intervention studies reviewed. Table 1 provides descriptions and properties of the established measures used to assess empathy and person-centeredness. Of the 18 total studies, only five measured empathy explicitly or personhood using either validated measures or modifications of an existing measure for the study. Two scales, the Jefferson Scale of Physician Empathy/ Health Professions (JSPE-HP; Fields et al., 2011; Hojat et al., 2001) and the Interpersonal Reactivity Index (IRI; Davis, 1980, 1983) have been used frequently in medical education research, but have been used less in examining empathy and caring for patients with dementia. The JSPE-HP scale is significantly correlated with conceptually relevant measures such as compassion (Hojat, Mangione, Kane, & Gonella, 2005). The IRI is not specific to health care and comprises four subscales, including (a) perspective taking, defined as the “tendency to spontaneously adopt the psychological point of view of others”; (b) fantasy, defined as “tendencies to transpose themselves imaginatively into the feelings and actions of fictitious characters”; (3) empathic concern, defined as “other-oriented feelings of sympathy and concern for unfortunate others”; and (4) personal distress, defined as “self-oriented feelings of personal anxiety and unease in tense interpersonal settings” (Davis, 1983, pp. 113–114). Passalacqua and Harwood (2012) measured empathy using a modified version of the IRI, only evaluating five items from the empathic concern and perspective taking subscales.
Other studies evaluated more implicit measures of empathy through person-centered beliefs or behaviors (Table 1). Four studies used the Approaches to Dementia Questionnaire (ADQ) to assess attitudes in staff working with persons with dementia. The ADQ includes an 11-item person-centered subscale. The ADQ does not explicitly measure empathy, but the “recognition of personhood” subscale measures the person-centeredness of care providers and includes items such as, “It is important to respond to people with dementia with empathy and understanding.” The Personhood in Dementia Questionnaire (PDQ) is a measure of person-centered attitudes toward persons with dementia in care providers working with this population (Hunter et al., 2013). This questionnaire addresses the personhood of persons with dementia, including their sense of purpose, ability or right to make choices, and feelings and emotions. In this way, it is an indirect measure of care providers' empathy.
Of the studies that measured communication skill competency, two used video recordings of patient–care provider interactions (Levy-Storms et al., 2016; Williams et al., 2017), whereas others collected data through direct observation (Burgio et al., 2001; Fritsch et al., 2016; Sprangers et al., 2015) or self-report (Passalacqua & Harwood, 2012). Another method of assessing the effectiveness of the intervention was through patient outcomes. Five studies measured resident emotions, such as depression or agitation (Eritz et al., 2016; Fritsch et al., 2016; McCallion et al., 1999; Sprangers et al., 2015; Teri et al., 2005). Five studies assessed the effect of the intervention on participant knowledge using mostly measures developed for the individual study (Table C).
Research Design and Study Quality. Eight studies used a pre-post design that analyzed before and after intervention effects (Adefila et al., 2016; Broughton et al., 2011; Cockbain et al., 2015; de Abreu et al., 2017; George et al., 2014; Levy-Storms et al., 2016; Passalacqua & Harwood, 2012; Ripich et al., 1995). Seven studies were RCTs with either the care provider (Beer et al., 2012; Hattink et al., 2015), nursing home resident (Eritz et al., 2016), or facility (Conway & Chenery, 2016; Fritsch et al., 2009; Sprangers et al., 2015; Teri et al., 2005; Williams et al., 2017) as the unit of analysis. Two studies used cluster randomization, in which groups of individuals were randomized by unit within facilities (Burgio et al., 2001; McCallion et al., 1999). Broughton et al. (2011) originally performed a cluster randomization by unit within the nursing homes but switched one of the control units to receive the intervention due to low recruitment at that facility. Control groups in five studies received no training (Broughton et al., 2011; Burgio et al., 2001; Conway & Chenery, 2016; Fritsch et al., 2009; Sprangers et al., 2015). Eritz et al. (2016) exposed control participants to resident medical histories rather than life histories; Teri et al. (2005) provided usual on-site training focused on the needs of older adults; and Beer et al. (2012) provided training on an unrelated topic (i.e., emergency medical services). Finally, three studies included a crossover design in which the control group received the intervention following completion by the experimental group (Hattink et al., 2015; McCallion et al., 1999; Williams et al., 2017).
Results of Individual Studies and Intervention Effect on Outcomes of Interest
Empathy. Of five studies that explicitly analyzed empathy or other qualities related to person-centeredness, three found significant improvements. Hattink et al. (2015) found formal dementia care providers significantly increased their empathic concern and perspective of others, and felt less personal distress (using subscales of the IRI) following an e-learning training, although no significant changes were seen in their attitudes toward persons with dementia (measured using the ADQ person-centered subscale). The simulation intervention evaluated by de Abreu et al. (2017) significantly enhanced students' attitudes, including the person-centeredness subscale in the ADQ (d = 0.48). Adefila et al. (2016) also found student empathy, operationalized as compassion, increased significantly after simulation (p < 0.001).
Three studies found no significant results, including one study assessing the global concept of empathy (using five items from the IRI) (Passalacqua & Harwood, 2012); a second study assessing attitudes toward persons with dementia, including recognition of personhood (Conway et al., 2015); and a third study using the JSPE-HPS (Eritz et al., 2016). However, Passalacqua and Harwood (2012) reported significant results following a communication training program using the ADQ (i.e., hope subscale, three items), although the “variable (hope) had reliability substantially below standard levels” (Passalacqua & Harwood, 2012, p. 442).
Communication With Empathy. Six studies focused on improvement of care provider communication with empathy or therapeutic communication skills for persons with dementia (Burgio et al., 2001; Fritsch et al., 2009; Levy-Storms et al., 2016; Passalacqua & Harwood, 2012; Sprangers et al., 2015; Williams et al., 2017) and all six reported positive outcomes. Four studies were experimental (Burgio et al., 2001; Fritsch et al., 2009; Sprangers et al., 20015; Williams et al., 2017) and a moderate global study quality rating was assigned in five of the six studies (Burgio et al., 2001; Levy-Storms et al., 2016; Passalacqua & Harwood, 2012; Sprangers et al. 2015; Williams et al., 2017).
The interventions varied and included a/an: communication skills training (Burgio et al., 2001; Passalacqua & Harwood, 2012); therapeutic communication program during mealtime (Levy-Storms et al., 2016) and morning care (Spranger et al., 2015); intervention to alert staff to elder-speak (Williams et al., 2017); and group story-telling program (Fritsch et al., 2009). The outcome measures varied across studies (Table C). Video recordings revealed CNAs' therapeutic communication behavior increased significantly post-intervention by a range of 27% to 48% (p < 0.05) (Levy-Storms et al., 2016), and Williams et al. (2017) found the Changing Talk (CHAT) intervention significantly improved the quality of communication in CNA–resident dyads. Specifically, elderspeak communication (i.e., use of terms of endearment, exaggerated voice intonation, looking away) declined significantly at 3-month follow up (34.6% vs. 12.2%; p = 0.016) (Williams et al., 2017).
Resident Outcomes. Only two of five studies that measured the effect of an intervention on resident symptoms including depression and agitation (McCallion et al., 1999; Teri et al., 2005) following a staff skills training program reported positive findings. In the McCallion et al. (1999) study, which evaluated the Nursing Assistant Communication Skills Program, depression symptoms and agitated behaviors were assessed. There was a significant decrease in all subscales of the Cornell Scale for Depression in Dementia in the treatment condition at 3 and 6 months, and a significant decrease was also noted in depression symptoms observed in the wait list control cross-over group. In addition, a significant decrease of physically nonaggressive behavior and aggressive behavior (using the Cohen-Mansfield Agitation Inventory [CMAI]) was demonstrated but for only 3 months following the intervention; the CMAI effects were not sustained at 6 months for two of the three subscales (McCallion et al., 1999). Teri et al. (2005) reported a significant decrease in assisted living residents' depression, problem behaviors, and anxiety symptoms 8 weeks after the Staff Training in Assisted Living Residences training program. This program addresses challenges of dementia care and tailors recommendations to the individual.
Other studies, however, did not report positive patient outcomes. Sprangers et al.'s (2015) evaluation of a communication skills training in a nursing home found no impact on the Dutch version of the CMAI. An intervention designed to familiarize nursing staff with residents' life histories found no impact on agitated behavior assessed with the CMAI (Eritz et al., 2016). A creative expression program resulted in an unanticipated increase in resident fear, anxiety, and sadness (Fritsch et al., 2009).