A best-evidence review of intervention studies for minimizing resistance-to-care behaviours for older adults with dementia in nursing homes


Correspondence to H.S. Kang:

e-mail: goodcare@cau.ac.kr



To conduct a best-evidence review of non-pharmacological interventions for resistance-to-care behaviours of nursing home residents with dementia in a personal-care context.


Resistance to care is a major source of staff burnout in nursing homes and it is also a safety issue for the staff.


Best-evidence review.

Data Sources

We searched for non-pharmacological intervention studies published from 1990–2012, written in English.

Review Methods

The search identified 19 intervention studies that examined the effects of interventions to reduce the resistance-to-care behaviours of nursing home residents with dementia in a personal-care context. These 19 papers met the quality assessment requirements of the critical appraisal criteria for experimental studies, which were published by the Joanna Briggs Institute.


Only three studies were RCTs and the rest were quasi-experimental. The sample size ranged from 7–127. Nine music interventions, such as pre-recorded music played to a group or playing a resident's preferred music, during his or her personal care, resulted in significant reductions in resistance-to-care behaviours. Resistance-to-care behaviours also were significantly reduced in three of four bathing interventions that focused on person-centred care. In the ability-focused interventions, only two out of five studies reported significant reductions in resistance-to-care behaviours.


Non-pharmacological interventions are options to consider to reduce resistance-to-care behaviours in older people with dementia, even though the evidence level is low, given the lack of alternatives. More randomized controlled trials are recommended to confirm the effects of non-pharmacological interventions during personal care.

Why is this review needed?

  • Evidence-based interventions for the management of resistance-to-care behaviours are crucial for both caregivers and older adults with dementia to maintain the quality of nursing care.
  • Evidence-based interventions will result in reduced caregiver burden and stress resulting from the resistant-to-care behaviours of nursing home residents with dementia.
  • Resistance-to-care behaviours are major barriers to personal nursing care, including feeding and bathing. There has not been a best-evidence review of the resistance-to-care behaviours of nursing home residents with dementia.

What are the key findings?

  • A low level of evidence suggests that using music during mealtime and shower/bathtime reduces the resistance-to-care behaviours of nursing home residents with dementia.
  • Some evidence suggests that a person-centred approach reduces resistance-to-care behaviours during bathing.
  • Weak evidence exists to suggest the use of an ability-focused approach during morning care reduces resistance-to-care behaviours.

How should the findings be used to influence policy, practice, research and education?

  • Person-centred care education is recommended to reduce resistance-to-care behaviours during personal care with older adults with dementia as a part of regular and continuing education sessions.
  • Playing pre-recorded relaxing music is recommended to reduce resistance-to-care behaviours during mealtime, if it is culturally appropriate.
  • Playing pre-recorded music based on the resident's preference might be useful to reduce resistance-to-care behaviours in older adults with dementia during personal care, if it is culturally appropriate.


In many countries, the proportion of older people with dementia is increasing steadily (Alzheimer's Association's 2012, Ikejima et al. 2012, Lopes et al. 2012). Many older people with dementia live in nursing homes (Moore et al. 2012) and most of them require assistance with personal care (Fjelltun et al. 2009). However, caregivers frequently experience resistance-to-care behaviours from nursing home residents, especially from older residents with dementia, in the context of providing daily personal care (Zeller et al. 2009, Ishii et al. 2012, Morgan et al. 2012).

Surveys in the Scandinavian countries revealed that the vast majority of nurses and caregivers had experienced residents' resistance-to-care behaviours. Seventy-nine per cent of 86 nursing home staff-members in Stockholm, Sweden (Forsell et al. 2011) and 96·9% of 494 nurses in 11 Norwegian nursing homes reported that they had experienced resistance-to-care behaviours of residents during personal care (Willumsen et al. 2012).

Resistance-to-care behaviours were described in several different ways, including the following: rejection of care (Ishii et al. 2012), resistant behaviour (Willumsen et al. 2012), resistiveness to care (Mahoney et al. 1999, Herman & Williams 2009), care-resistant behaviour (Jablonski et al. 2011) and resistance to care (Potts et al. 1996, Werner et al. 2002). Potts et al. (1996) defined ‘resisting care’ as ‘any patient behaviour which prevents or interferes with the nurse performing or assisting with activities of daily living (ADLs) for the patient, including bathing, toileting and grooming (p. 11).’ For this review, the term ‘resistance-to-care’ will be used to describe these behaviours to maintain consistency.

In the past decade, the interpretation of resistance-to-care behaviours has changed from being considered symptoms of dysfunctional cognitive status to being considered meaningful behaviour requiring validation of unmet needs (Ayalon et al. 2006, Spira & Edelstein 2006). This shift has been facilitated by the person-centred care approach (Kitwood 1997), which sees the personhood of individuals with dementia in all aspects of care. In this approach, dementia-related behaviours, such as resistance-to-care behaviours are seen as ‘ways to communicate unmet basic needs, such as hunger or pain’ (Edvardsson et al. 2008, p. 363).

Resistance-to-care behaviours are a major barrier to the safety and comfort of people with dementia in the context of personal care (Forsell et al. 2011, Willumsen et al. 2012). Furthermore, from the perspective of caregivers, resistance-to-care behaviours can cause distress and can lead to abusive behaviours by caregivers (Volicer et al. 2009) or the need for physical and pharmacological restraints (Werner et al. 2002). Thus, the effective management and prevention of such behaviours are important global issues.

Several systematic literature reviews related to resistance-to-care behaviours have been conducted, including the following: oral hygiene care for adults with dementia in residential older-care facilities (Chalmers & Pearson 2005); the aggressive behaviour of nursing home residents towards caregivers (Zeller et al. 2009); the therapeutic effects of music and singing for older people (Skingley & Vella-Burrows 2010); the effect of music at mealtimes for people with dementia (Johnson & Taylor 2011); feeding-related nursing care (Watson & Green 2006); and interventions used to promote oral nutritional intake among older people with dementia and feeding difficulties (Watson & Green 2006). However, these reviews were not solely focused on resistance-to-care behaviours in the context of daily personal care, or they focused solely on a specific care situation, such as oral care, bathing, or mealtime. As a result, questions were raised about whether interventions should be offered differently for each of the care activities during personal care.

Thus, the aim of the present review was to examine what interventions effectively manage or reduce the resistance-to-care behaviours of nursing home residents with dementia. This review focused on the effective interventions for overall daily personal care. In this review, we expected to be able to provide practical recommendations concerning effective nursing care for overall daily care and for each daily care activity for the management of resistance-to-care behaviours of older adults with dementia, living in nursing homes.

For the purpose of this review, the term ‘dementia’ is considered from both medical and social models (National Collaborating Centre for Mental Health 2007). ‘Dementia’ was interpreted as both a global cognitive impairment associated with impairment in functional abilities and a behavioural and psychiatric disturbance, and a form of disability caused by social exclusion and inappropriate treatments provided by older care providers and society to older people with dementia (National Collaborating Centre for Mental Health 2007).

The review


This best-evidence review aimed to determine: (1) the effectiveness of interventions intended to reduce the frequency and intensity of resistance-to-care behaviours during assisted personal-care activities (i.e. mealtime assistance, toileting, oral care, morning care and showering/bathing); and (2) the effectiveness of interventions to reduce the overall frequency and intensity of resistance-to-care behaviours in the daily, personal-care activities of nursing home residents with dementia.


This best-evidence review followed the guidelines for systematic reviews of quantitative evidence proposed by the Joanna Briggs Institute (JBI) (Joanna Briggs Institute 2011).

Studies were considered for inclusion that: (1) examined any non-pharmacological intervention to reduce the resistance-to-care behaviours of people with dementia over the age of 55, who were living in a residential-care setting; (2) used outcome measures of resistance-to-care frequency and intensity during personal care; and (3) were randomized controlled trials (RCT), or quasi-experimental studies, published in English. Studies with outcomes related to caregivers were not considered (e.g. the competency or the perceptions of the caregivers).

Search methods

The databases searched were CINAHL, MEDLINE, Embase, PsycINFO, Cochrane Clinical Trials, SCOPUS, ProQuest (dissertations), Web of Knowledge, Mosby's Nursing Consult and Health Science/Nursing Academic Edition. Studies published from 1990–2012 in English were included in the review. The year 1990 was chosen because systematic reviews on non-pharmacological interventions in dementia patients found no interventions reported prior to 1990 (Chalmers & Pearson 2005, Watson & Green 2006, Zeller et al. 2009, Skingley & Vella-Burrows 2010). Furthermore, Kitwood's seminal work on person-centred dementia care was published in 1997 and studies related to person-centred care of dementia started appearing around that year in the major databases. The reference lists of all of the included articles were individually searched for additional studies. The following key words were used: resist*, reject*, resistance, resistiveness, refusal, agitation, disruptive behaviour, dementia, care and nursing care. Studies were considered for review if the study population had been diagnosed based on standardized and validated assessment criteria, such as the Mini-Mental Status Examination. However, given the absence of high-quality RCTs, studies using a diagnosis of dementia were accepted for a review.

Search outcome

The database search yielded 1262 papers. Based on the titles and abstracts, 103 studies were selected for a detailed review by the first and second authors. The reference lists of studies selected for quality appraisal were hand searched individually and two additional studies were identified.

Quality appraisal

The 19 selected studies were examined for potential bias by two independent reviewers prior to inclusion in the review. The critical appraisal checklist for experimental studies, published by the JBI (Joanna Briggs Institute 2011), was adopted. This checklist contains 10 criteria for detecting selection bias, performance bias, detection bias and attribution bias. Case studies were excluded because they have no external validity. Because of the limited number of high-quality intervention studies, all 19 studies were included in the review regardless of their quality.

Data abstraction

The data from the 19 included studies were abstracted using a tabular table (Table S1). The first and second authors independently abstracted the data from the included studies; any disagreement between the reviewers was resolved through consensus.


Originally, the possibilities of pooling data were considered. However, because of the heterogeneity of types of interventions and study designs, the data were analysed narratively. A flow diagram of the study selection process is presented in Figure 1.

Figure 1.

Flow diagram of the study selection process.


Study design and sample

All 19 studies examined the effectiveness of intervention(s) to minimize the resistance-to-care behaviours of older adults with dementia in the context of personal care (Table 1). Of the 19 studies reviewed, 16 were quasi-experimental studies and three were randomized controlled trials (RCTs). The random allocation of participants to treatment or control groups, a cross-over design and self-controls were characteristics of the RCTs. Most of the quasi-experimental studies used a pre–post design or a time-series design that repeated pre–post measures, or measured outcomes in the postintervention phase. One study reported using a time-series design (Richeson & Neil 2004); however, in actuality, the study used a pre–post design. Detailed information about all of the included studies is provided in Table 1.

Table 1. Summary of the characteristics of the reviewed studies
StudyStudy design #(screening for agitated behaviours)Type of interventionDuration of intervention; Number of observationsType of observation and observers & MeasurementSample & Setting (Mean)Significant improvement
  1. *P < 0·05; **< 0·01; ***< 0·001.

  2. a

    Statistical tests were conducted by the reviewers.

  3. Study design: A, Non-intervention phase; B, Intervention phase.

Meal time care
Richeson and Neil (2004)a



Music at dinner (3 different relaxing music CD)

4 days each in 2 weeks:

4 for A, 4 for B

Direct observation by the researchers

● Modified CMAI


= 27, mean age: 87 years.

Number of agitated behaviours*
Hicks-Moore (2005)aTime study approach (ABAB)Music (relaxed with the Classics) at dinner1 week each, 4 weeks in total:14 for A, 14 for B

Direct observation by 1 researcher

● Modified CMAI

Specialized care unit, Canada = 30, age: range 70–101 years (82 years)Total agitated behaviours***
Ho et al. (2011)quasi-experimental (BA)#Music (composed by the researcher) at lunch and dinner4 weeks for B; 2 weeks for A: 24-hour record for 7 days/week

Direct observation by 6 trained staff nurse


NH, Taiwan

= 22, age: 62–91 (77 ± 8 years), MMSE: 6–22 (11 ± 5)

Mean total agitated scores***; 4 subscale agitated scores***
Denney (1997)aTime study approach (ABAB)#Music (relax with the classic) at lunch7 days each; 2 for A, 2 for B

Direct observation by the author

● Modified CMAI

Aged care facility, Canada

= 9, age: 65–84 (75 ± 6 years), MMSE: (3 ± 5)

Total agitated behaviours**; physically non-aggressive behaviour**
Chang et al. (2010)Time study approach (ABABABAB)#Music (nature music by broadcast network) at lunch7 days each, 8 weeks in total; daily between 09:00-21:00

Direct observation by 6 student nurses or social workers

● type and frequency of problem behaviour

NH, Taiwan

= 41, age: 69–94 (82 ± 6 years), MMSE: 0–21 (11 ± 7)

Physically aggressive behaviours*, verbally aggressive behaviours*, total CMAI score*
Goddaer and Abraham (1994)Time study approach (ABAB)Music (relaxing music tape) at lunch1 week each; daily (lunch time only)

Direct observation by 1 research assistant

● Modified CMAI.

2 NHs, Belgium

= 29, age: 67–93 (81 ± 7 years), MMSE: 0–17

Total agitated behaviours***; verbally agitated behaviours**; physically non-aggressive behaviours*
Whall et al. (1997)Pre–post with control (AB)#Simulated natural environment (e.g. bird songs & pictures, & food)1 bath (A), 2 baths (B); 1 week apart; 1 for A, 2 for B; 3 for control

Direct observation by 2 expert research assistants

● Modified CMAI

5 NHs, USA

= 31 (15 in intervention group, 16 in the usual care, MMSE<5

Agitated behaviour (mean CMAI score)**
Thomas et al. (1997)Time study approach (ABA)Music: selected by the family member played in the bathroom3 showers at each phase; 6 for A, 3 for B

Direct observation by certified nurse assistants

● Modified CMAI

Nursing facility for aged, USA = 14, age range (69–86 years)Frequency of aggressive behaviour*
Clark et al. (1998)quasi-experimental#Music: preferred by the participant played during bathing

2 weeks each:

20 sessions per subject

Direct observation by 4 research assistants

● a list of behaviours (hitting, biting, yelling, etc.).


= 18, age: 55–95 (82 ± 10 years).

● MMSE score: (1–22, 10 ± 7),

Total number of aggressive behaviours*; hitting behaviours*
Dunn et al. (2002)RCT with cross-overThermal bed bath compared with tab, ‘person centered’4 consecutive sessions each; 4 for thermal baths, 4 for tab

Direct observation by 2 authors

● CMAI and Ryden Aggression Scale

Urban continuing care facility, Canada

= 16, age: 67–93 (81 ± 7 years), BCRS: 3·8–7

Sum of 14 agitated behaviours**
Mickus (2002)Pre–post (AB) #Agitation prevention: ‘person centered’1 for A, 1 for B

Direct observation by 1 research assistant

● 5 subscales of Neuropsychiatric Inventory during bathing


= 23, Mean age (77 ± 20 years), MMSE (M-6, range 0–24)

Irritability*; anxiety*
Hoeffer et al. (1997)Pre–post (AB)#Person-focused bathing care and bedside consultation2–4 weeks to establish person-focused bathing; 3–6 baths for A, 8 for B

Direct observation by 3 nursing assistants

● Revised version of the Ryden Aggression Scale

Intermediate care facility, USA = 10, age: 86–98 (90 years), MMSE ≤ 21 (5·7)Physical aggression*; verbal aggression*; being upset*
Sloane et al. (2004)RCT (cross-over of 2 interventions with control)#Person-centred showering & towel bath6 weeks each; not stated

Video recording; coding by blinded research assistants

● Care Recipient Behavior Assessment (modified CMAI), modified discomfort scale

15 NH, USA

= 69, age:(Control: 87 ± 6 years; Intervention: 86 ± 9), MMSE: (C: 2 ± 4; I: 2 ± 4),

Any agitation & aggression*; any physical aggression**; discomfort scores***
Morning care
Hammar et al. (2011)Pre–post (AB)Music: caregivers' singing old songs or hymns4 weeks each; 4 for A, 4 for B; 10–20 minutes observation

Video recording; coding by 2 researchers

● Observed Emotion Rating Scale; Restiveness To Care Scale

2 NHs, Sweden

= 10, age: 66–92 years, MMSE: mean = 3, range: 0–12; 10 caregivers

Behaviours (pull away*, grab object *, adduct*); positive emotions (pleasure*, general alertness*)
Rogers et al. (1999)Pre–post (AB)Rehabilitation (behavioural and occupational therapy techniques): ‘ability focused’5 days for skill elicitation & 15 days for habit training follow-up; 2 days for each phase

Direct observation by a research associate

● type, number and duration of sessions on real time

5 NHs, USA

= 84, mean age: 82 ± 6 years), MMSE (M = 6)

Disruptive behaviour**; proportion of time engaged in non-assisted and assisted dressing*
Sidani et al. (2012)Pre–post (AB)Ability-focused morning care (AFMC)posttest observation <1 month after education session; 2 for A, 2 for B

Direct observation by the staff nurse and/or a research assistant

● Pittsburgh Agitation Scale; Modified Interaction Behavior Measure; list of AFMC

8 long-term care facilities, Canada

= 65; age: 58–100 (83 ± 8 years); MMSE: 0–25 (6 ± 8)

Wells et al. (2000)Pre–post (AB)Caregiver education programme on social abilities and self-care abilities: ‘ability focused’baseline (? days), follow-up at 3 and 6 months; not stated

Direct observation by a research assistant

● Modified Interaction Behavior Measure; Pittsburgh Agitation Scale; London Psychogeriatric Rating Scale (LPRS)

4 cognitive support units, NH, USA

= 40, (20 in intervention in 1 unit, 20 in control in 3 units)

Agitation*; Pittsburgh Agitation Scale*; personal attending*; calm/functional behaviours*; social function*; LPRS*
Mouth care
Jablonski et al. (2011)Pre–post (AB)#Mouth care (Managing Oral Hygiene Using Threat Reduction: ‘ability focused’3 days for baseline; twice daily for a 2-week intervention; 6 for A, 28 for B

Direct observation by the research team

● Modified Restiveness To Care Scale; modified Restiveness to Care Scale for oral health


= 7, age (82 ± 4 years), (GDS: 6 ± 0·9)

Care-resistant behaviours (P = 0·06); improved oral health**
Daily basic care
Beck et al. 2002,RCT with cross-over#Interventions: (1) ADL, (2) psychosocial activity, (3) combined; (4) placebo, & (5) no intervention: ‘ability focused’7-week intervention (5 days/week); at 1-month & 2-month follow-up; not stated

Video recorded; coded by 6 trained raters; Disruptive Behavior Scale/8-hour shift recorded by nursing assistant:

● Disruptive Behavior Scale; Observable Displays of Affect Scale; Apparent Affect Rating Scale

7 aged care facilities in 2 states, USA

= 127, age: (83 ± 7·4 years), MMSE: (10 ± 6·3),

Positive facial expressions**; positive body posture**

All studies reported the level of dementia severity in their sample, except two studies, which only stated the diagnosis was dementia (Richeson & Neil 2004, Hicks-Moore 2005). Of 17 studies that reported severity, MMSE was used in 13 studies to estimate the severity of the dementia of the participants. The other scales used were the Global Deterioration Scale (GDS) (Thomas et al. 1997, Jablonski et al. 2011), the brief cognitive rating scale (BCRS) (Dunn et al. 2002) and the Minimum Data Set Cognitive Score (MDS-COGS) (Sloane et al. 2004). Most of the studies classified their participants as having moderate or severe levels of dementia.


The main outcomes that were measured by the studies varied, including disruptive behaviour, problem behaviour, agitation, aggression and resistance-to-care. A variety of scales were used to measure the main outcomes. The most frequently used scale was a modified Cohen-Mansfield Agitation Inventory (CMAI) to measure the presence or absence of agitated behaviour (Table 1). The CMAI is composed of four categories: physically non-aggressive behaviour, physically aggressive behaviour, verbally non-aggressive behaviour and verbally aggressive behaviour. No other scale was used by more than two of the studies (Table 1).

Reliability and/or validity were reported in most of the studies. However, the validity and reliability were not reported in the studies that used self-developed tools. Most of the studies measured only the disruptive behaviours of the participants for intervention outcomes, except for 4 studies that also measured emotion.

Methodological quality

Overall, most of the studies had a high risk of bias because they used a quasi-experimental design with a small sample size. The risk of bias assessment results for each of the included studies are presented in Table 1.

The sample sizes ranged from seven (Jablonski et al. 2011) to 127 (Beck et al. 2002). The number of facilities that participated in each study varied from 1–15. The majority of the studies were based on only one facility. Furthermore, the majority of participants were women with moderate-to-severe dementia, with a mean age of 80–90 years.

A history of resistance-to-care or related behaviours as a part of the eligibility criteria would affect the incidence of resistance to care; this was mentioned in 10 studies (Table 1). Of these 10, seven studies mentioned the prevalence of resistance-to-care behaviour that occurred during care. Resistance-to-care frequency was described non-quantitatively (e.g. regularly, consistent, frequently). Only two studies quantified the frequency of resistance to care as ‘at least half the time’ (Whall et al. 1997) or ‘during two of four baths’ (Hoeffer et al. 1997).

Direct observation was used in 16 studies; only three studies used video recording (Table 1). The number of observations per participant was not well described in most of these studies. The number of observation per participant varied from two for the entire study period (Mickus 2002) to 34 (Jablonski et al. 2011). Moreover, in three studies, the number of observations in the pre-intervention period was fewer than that in the intervention period (Thomas et al. 1997, Whall et al. 1997, Jablonski et al. 2011).

Inter-rater reliability was reported in about half of the studies (Goddaer & Abraham 1994, Richeson & Neil 2004). In seven studies, the training for interobserver agreement was provided without a minimum agreement rate (Thomas et al. 1997, Beck et al. 2002, Richeson & Neil 2004, Chang et al. 2010, Hammar et al. 2011, Ho et al. 2011, Sidani et al. 2012). For this review, the interventions were assigned to one of two categories: (1) interventions by environmental control; and (2) educational interventions for daily care.

Interventions with environmental control

Ten interventions involved environmental control. Of these, six were mealtime music interventions, three pertained to bathing care and one was a music intervention for morning care (Table 1). Pre-recorded music was used for the mealtime interventions, predominantly relaxing classics. The music selection for the bathing interventions was based on the patient's preference or a family member's choice (Thomas et al. 1997, Clark et al. 1998). The choice of music for singing by caregivers during morning care was old songs or hymns with which the patients were familiar (Hammar et al. 2011). One non-musical intervention used a simulated natural environment in the bathroom. This intervention had four components: pictures, sounds of nature, offering of sweets and caregiver interaction regarding the natural environment (Whall et al. 1997).

With regard to data collection, direct observation was used to monitor resistance to care in nine studies and in three studies the staff observed the patients (Table 1). Video recording was used in the one study where the caregivers sang songs during morning care (Table 1). For the most part, the duration of the observations for mealtime music interventions was limited to mealtime. However, in two music-intervention studies, the observation period was longer. One was a lunchtime intervention with an observation period of 12 hours (9 AM–9 PM) (Chang et al. 2010); the other was for both lunch and dinner, with a 24-hour observation period (Ho et al. 2011). Another issue was the observer-to-participation ratio, which was difficult to calculate because of the limited information that was provided. In one study, one researcher observed 30 participants (Hicks-Moore 2005). The low observer-to-participant ratio should be examined, although it depends on the incidence of resistance-to-care.

Three mealtime music-intervention studies used a time study approach with repeated interventions (Goddaer & Abraham 1994, Denney 1997, Chang et al. 2010). In these studies, the music intervention was reported to have a carry-over effect. A carry-over effect was also observed in the music intervention for bathing (Thomas et al. 1997). Although we could not determine the study design of one of the music- intervention studies (Clark et al. 1998), a music intervention was reported to have been added to the bathing care routine.

The study participants’ reactions to music were reported in two of the studies. Caregivers noted that the patients’ mood improved during the music intervention and some patients smiled and danced (Clark et al. 1998). One other study also reported similar reactions to a music intervention, stating that some of the residents swayed and clapped their hands in tune to the music (Hicks-Moore 2005).

Of 10 studies reporting on environmental control interventions, seven reported statistically significant reductions of aggressive behaviours, despite differences in the measurement methods and observation intensity between the studies (Table 1). The remaining three studies presented only descriptive data (Denney 1997, Richeson & Neil 2004, Hicks-Moore 2005), so the authors of this review calculated the statistical significance in these studies. The incidence rate of resistance-to-care behaviour per observed unit in the pre-intervention period was treated as the expected rate and incidence rate in the postintervention period was treated as the observed rate. When the statistical significance of this standard incidence ratio was tested using the chi-square test (Checkoway et al. 1989), all three studies showed statistically significant reductions in resistant-to-care behaviours (Table 1). A few studies reported substantial fluctuations in the occurrence of agitation in some participants (Hicks-Moore 2005, Chang et al. 2010).

Educational programmes for caregivers

Nine interventions included a caregiver education component for reducing resistance to care during daily care. Four studies aimed to improve the caregivers’ person-centred bath skills; specifically, to respect the patient's privacy and to improve their comfort during bathing. Five studies used ability-focused interventions that were designed to improve caregivers’ assessment skills, specifically, to improve the caregivers’ ability to evaluate the patients’ abilities and improve their care skills (Table 1).

A person-centred educational programme for bathing

A person-centred educational programme was applied only to bathing care in four studies. Only 1 RCT used a cluster-randomized design to evaluate person-centred showering and towel bathing among 67 participants (Sloane et al. 2004). Person-centred showering (which declined by 53% of baseline, < 0·001) and towel bathing (which declined by 60% of baseline, < 0·001) were associated with significantly lower resistance to care compared with a non-person-centred shower control group. In this study, intensive one-on-one training was offered to the nurse's aides. Approximately 8 hours per nursing home participant was required for training and developing person-centred bathing (Sloane et al. 2004). Video recording was used for evaluation.

A brief report described the effect of a 30-minute interactive seminar for nurse's aides on the topic of basic person-centred bathing care (Mickus 2002). This pragmatic education programme resulted in a significant reduction of irritability and anxiety but not agitation during bathing in a sample of 23 residents. However, observations were made only twice: once in the pre-intervention and once in the postintervention period. The other two studies were small-scale studies with sample sizes ≤16, which reported significant reductions in agitated behaviours (Table 1).

An ability-focused educational intervention for daily care and morning care

The ability-focused approach was developed under the influence of person-centred care (Sidani et al. 2012), which embraces the personhood and dignity of people with dementia. The ability-focused approach focuses on maximizing and maintaining a person's retained ability in physical and psychosocial functions by creating a supportive environment (Beck et al. 2002, Sidani et al. 2012).

An ability-focused educational programme for caregivers was employed in one RCT and three pre–post studies with sample sizes ranging from 40–127 and one pilot study of an oral-care intervention. The RCT used a cross-over design with five arms (Beck et al. 2002); three types of interventions to reduce resistance-to-care during daily care were evaluated in 127 nursing home residents in seven facilities. The first arm was an activities-of-daily-living (ADL) intervention that was conducted during bathing, grooming, dressing and lunch times, based on a behavioural model that focused on residents’ psychosocial needs. The second arm used a psychosocial approach that focused on the need for communication, self-esteem and a sense of safety and security. The third arm used both of these methods. The study used a placebo control and a no-intervention group. The three types of interventions did not significantly reduce the resistance-to-care behaviours.

It is worthwhile to briefly describe the three pre–post ability-focused morning care studies. In a study by Sidani et al., 1·5–2 hours educational sessions were offered to the caregivers along with consultation to improve ability-focused morning care in five nursing homes (Sidani et al. 2012). This intervention did not result in statistically significant outcomes, such as agitation or participation in morning care, for the 65 residents.

The other two educational programmes for caregivers had smaller sample sizes and the intervention was conducted at one institution. One aimed to improve social and self-care abilities in 40 residents and showed significant declines in agitation (= 4·06, < 0·05) (Wells et al. 2000). The other study, with a larger sample size, did not include an educational programme for caregivers. This study was implemented by a research therapist who used a behavioural and occupational therapy technique to assess and train 84 participants with dementia (Rogers et al. 1999). The intervention resulted in a significant increase in ADL participation and a decrease in disruptive behaviour (= 2·74, < 0·01).

Fluctuations in the incidence of agitation were reported in two studies (Wells et al. 2000, Jablonski et al. 2011). The level of agitation fluctuated across three points in time for both the experimental and the control groups. A between-group difference in mean agitation scores was evident only at a 6-month follow-up (= 2·12, = 0·041) (Wells et al. 2000). Thus, the significant decline in the agitation scores might not have been caused by the intervention. In a pilot mouth-care study, substantial fluctuations in the resistance-to-care scores of seven participants were reported during the baseline and the intervention periods (Jablonski et al. 2011).


This review examined studies on non-pharmacological interventions for minimizing the resistance-to-care behaviours of nursing home residents with dementia during assisted personal care. Although nearly all the studies reported reduced resistance-to-care behaviours, the quality of evidence was generally low because of the use of quasi-experimental designs and measurement problems.

One of the major issues this review identified was related to the measurement of resistance-to-care behaviour. In some studies, the care staff who implemented the intervention also collected the data, which might have compromised objectivity. Another problem was the substantial fluctuation in the occurrence of resistance-to-care reported in several studies. This means that a small number of observations would decrease statistical power. Furthermore, in several studies, the number of observations during the pre-intervention period differed from that in the postintervention period. This suggests a between-time-period difference in the probability of observing resistance-to-care behaviours. There should be an equal number of observations during the pre- and postintervention phases. Alternatively, resistance to care could be expressed as the incidence of resistance to care per week so that those measurements could be compared.

Participation bias was another major source of bias. The nursing homes and caregivers were not randomly selected. Those who were willing and able tended to be selected for the study or volunteered to participate. Participants with a history of resistance to care would increase resistance-to-care incidence and the resistance-to-care inclusion criterion would increase the statistical power to detect an intervention effect. However, we did not observe this phenomenon. Participation bias limits the generalizability of these intensive educational programmes.

The studies that examined two types of interventions are discussed in the light of these limitations. One type of intervention involved environmental control and the other intervention involved caregiver education.

Environmental control

Playing music was the most commonly reported intervention used to reduce resistance-to-care behaviours during personal-care activities for residents with dementia. Although the level of evidence is weak, the uniformity of the direction of the effect suggests that using music can reduce resistance-to-care behaviours during mealtime, bath/shower time and possibly in other personal-care contexts. Moreover, as there was no report of adverse events, this suggests that music is a safe intervention to manage resistance-to-care behaviours. Unlike formal music interventions that require the involvement of music therapists, the music interventions identified in this review (i.e. playing a compact disc and caregiver's singing) can easily be implemented by the staff. These methods do not require lengthy staff training and/or new employment of music therapists. Consequently, additional costs are not a factor. If it is culturally appropriate to play music during care, music familiar to the residents’ generation is an option to consider. It could serve as reminiscence therapy and promote conversation among the residents (Hicks-Moore 2005).

Mealtime music interventions might have an impact on resistance-to-care behaviours all day. In two interventions, observations were made during day-shift hours or a 24-hours period, which implies that mealtime music interventions might have affected the occurrence of resistance-to-care behaviours during the daytime or a 24-hours period. It is possible that the mealtime intervention had an impact on the rest of the day. To examine this effect, observations should be conducted separately during mealtime and the rest of the day, rather than treating a day as one time-block of observation. However, to increase the generalizability of the evidence, it is necessary to clarify the best volume, type of music, timing and length of music play, as well as clarifying variations in the effects for individuals with different types of dementia and different severity of dementia.

Educational programmes for caregivers

Person-centred care for bathing

A person-centred care approach was employed only in the bathing care intervention. All four studies, including two pilot studies, demonstrated a significant effect of person-centred care. In particular, one RCT conducted by Sloane and associates provided relatively strong evidence to support a person-centred approach for bathing assistance (Sloane et al. 2004), with no incidence of undesirable effects reported. However, bathing activity is culturally sensitive. All four studies were conducted in Western countries and the applicability of this evidence to other cultures needs further clarification. The variation of effects on different dementia types also requires greater clarification to increase its generalizability. Furthermore, educational programmes tend to entail substantial costs and staff time. Although the validity of the study was low, Mickus (2002) reported a simple in-service training that provided a pragmatic approach to implementing person-centred care. Even though the level of evidence for person-centred care is low, it is recommended for persons with cognitive impairment. A cost-effective educational programme should be developed.

An educational programme with an ability-focused approach for daily basic care or morning care

The effects of the ability-focused approach varied. No effect on resistance-to-care behaviours was reported in two large, multiple-site studies (Beck et al. 2002, Sidani et al. 2012). On the other hand, the smaller study (Wells et al. 2000) reported a significant decrease in resistance-to-care behaviours. These conflicting findings may be caused, in part, by the difficulty of implementing an ability-focused educational programme for caregivers in the larger studies.

The use of different scales and measurement methods may have contributed to the discrepant findings. Theoretically, ability-focused care is intended to promote residents’ involvement in their care on the basis of their individual ability; thus, promoting the resident's engagement may reduce disruptive behaviours (Bharwani et al. 2012). Nevertheless, the description of specific educational programmes was mostly too brief to compare them among studies and contributing factors to resistance to care, such as environmental factors and staffing levels, were not available for comparison.


The conceptual framework of resistance to care and resistance-to-care measurement were not reviewed here; thus, they require a separate review. Refinements of the conceptual framework will facilitate resistance-to-care research as hypotheses become testable. In addition, the selection of a resistance-to-care measurement scale would have an effect on the incidence of resistance-to-care behaviours. This effect could be tested by analysing video-recorded data in a study that used multiple resistance-to-care scales. The differences in the type and incidence of resistance-to-care between the scales could be examined.

The study only included studies published in English, which may have increased the risk of publication bias. Future systematic reviews should include studies written in languages other than English. The use of diverse cultural settings would increase the sample size and generalizability of the evidence.

Implications for research

The present review identified several areas where further research is needed. Although only behaviour was examined in most of the studies, some also examined the participants’ emotional states (Chang et al. 2010) and the possible effect of the intervention on caregiver stress (Clark et al. 1998). Caregivers’ confidence level about caregiving should also be examined because the intervention outcome could be affected by their skill, expertise, or confidence. In addition, the working environment of caregivers and the living environment of residents have to be examined.

Inadequate staffing and workload would affect the stress level of caregivers and residential, environmental factors, such as noise levels and personal space issues, are also known to be contributing factors to resistance-to-care behaviours (Rasin & Barrick 2004, Kable et al. 2013).

This review identified several terms such as ‘agitation,’ ‘aggression,’ and ‘disruptive behaviours’ that are used to describe the resistance-to-care phenomena. A standardized terminology should be developed, because the use of a common terminology would help improve communication between researchers and caregivers and build solid and consistent nursing knowledge.

Although the identification of risk factors for resistance-to-care behaviours was beyond the scope of this review, several studies reported triggers of such behaviours, including water spray and caregivers’ impersonal attitudes (Rasin & Barrick 2004, Rader et al. 2006, Cohen-Mansfield & Parpura-Gill 2007). Therefore, a comprehensive approach is recommended when providing assisted bathing. This comprehensive approach should consider risk factors and effective interventions for managing resistance-to-care behaviours.

Implications for practice

Regular and continuing education sessions should include training in the interventions that effectively reduce resistance-to-care behaviours during personal care. Moreover, caregivers should be aware of the inclusion criteria for each of the interventions because some are limited to residents with certain intact functions (e.g. cognitive or physical). For example, caregivers should ensure that residents with hearing difficulties are wearing their hearing aids before playing music.

Furthermore, Sloane et al. (2004) reported that average bathing time increased significantly as a result of person-centred showering. This finding is consistent with a study that reported that more time is needed for residents who have greater levels of cognitive impairment (Kobayashi & Yamamoto 2004). Employers should realize that the use of person-centred care will lead to more time being devoted to caring for residents with severe dementia. However, the reduction of resistance-to-care behaviour during care is expected to reduce staff stress and burnout. Institutional support for caregivers is necessary to enable the flexibility of time spent for assisted bathing.


We reviewed the effects of non-pharmacological interventions on the resistance-to-care behaviours of residents with dementia in a personal-care context. Interventions were mostly targeted for mealtime care, bathing and morning care. The level of recommendation for the non-pharmacological interventions for resistance-to-care behaviours is low because of problems in study design, measurement of resistance to care and the evaluation of interventions, regardless of the type of care. However, most of the studies showed significant reductions in resistant-to-care behaviours. Providing culturally sensitive, person-centred care on the basis of individual preferences and abilities is a basic principle for personal care considering the alternative of non-person-centred care. Future research needs to overcome the problem of the measurement and evaluation of the effects of non-pharmacological interventions during personal care.


This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest

No conflict of interest has been declared by the authors.

Author contributions

All authors have agreed on the final version and meet at least one of the following criteria [recommended by the ICMJE (http://www.icmje.org/ethical_1author.html)]:

  • substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data;
  • drafting the article or revising it critically for important intellectual content.