Interventions to improve the oral health of people with dementia or cognitive impairment: A review of the literature

Abstract

Objective

Oral diseases and conditions are prevalent among older people with dementia and cognitive impairment. While many interventions have been advocated for use in this population, evidence for their effectiveness is unclear. Our objective was to review systematically the content and effectiveness of interventions and implementation strategies used to improve or maintain the oral health of people with dementia or cognitive impairment.

Methods

Original studies published in English at any time until January 2015 were identified through electronic searches of the Medline, Embase, CINAHL, Scopus and Cochrane databases and hand searches of eligible studies and relevant reviews. Two investigators independently abstracted study characteristics and assessed the methodological quality of eligible studies. Results were presented as a narrative review because significant heterogeneity among included studies precluded a meta-analysis.

Results

The 18 included studies varied considerably in terms of size, scope and focus. Only two studies were identified that had been designed specifically for and examined exclusively in people with dementia or cognitive impairment. All studies were in residential care; none was population-based. While several studies reported positive effects, a number of methodological weaknesses were identified and the overall quality of included studies was poor. The specific outcomes targeted varied across studies but most studies focused almost exclusively on proximal clinical oral health outcomes such as levels of dental or denture plaque. Attempts to measure intervention integrity were limited and there was usually little or no effort to evaluate intervention effects over a sustained period.

Conclusion

There is a lack of high quality evidence to support the effectiveness of oral health interventions and implementation strategies for older people with dementia or cognitive impairment. More rigorous, large scale research is needed in this area. Recommendations are provided to improve the overall quality of evaluation in this area. Emphasis must be placed on developing evidence-based, achievable and sustainable oral health strategies if the needs of people with dementia and cognitive impairment are to be met into the future.

Introduction

Oral diseases are highly prevalent in the older population (1, 2). This vulnerability is associated with older adults' greater physical frailty, medical comorbidities, polypharmacy, cognitive impairment and functional dependence (3). Poor oral health can lead to pain, impaired general health and reduced quality of life. Oral conditions and dental pain affect chewing and swallowing, nutritional intake, mood, behaviour, selfesteem and social interaction (4). There is also a growing interest in the relation between oral and general health, as poor oral health is associated with a range of adverse health outcomes including cardiovascular disease, diabetes and respiratory disease (5).

Oral diseases and conditions are more prevalent among certain subgroups of older people, such as people with dementia and cognitive impairment (6, 7). Comorbidities and barriers to care are observed at even higher rates in these individuals, who are more functionally dependent, medically and nutritionally compromised and cognitively impaired than their cognitively intact counterparts (8). Oral health maintenance in older adults with dementia represents a unique challenge for researchers and clinicians, as they are more likely to reside in long-term care, depend on caregivers for routine oral care and display greater resistance to oral care provision (6).

There is evidence to support the use of a range of preventive oral hygiene care strategies in older adults and adults with ‘special needs' (9). However, the most recent systematic review of oral hygiene care for people with dementia, conducted over a decade ago, found that few of these strategies had been directly tested in this population (9). The oral health of people with dementia and cognitive impairment may also benefit from oral health care training for staff and caregivers, to facilitate better implementation of established guidelines in residential aged care and in the community. The majority of residents in long-term care require some assistance with activities of daily living and people with dementia are among the most functionally dependent. Many implementation strategies have been designed and evaluated to enhance the ability of nurses and caregivers to improve the oral health of older people residing in residential care (10, 11). While many of these strategies improve knowledge among caregivers, they have not been found to have a consistently positive effect on oral health (11). Due to methodological issues, it has not been possible to identify which strategies are most effective in improving oral health of people in residential aged care. It is also unclear which interventions, if any, would benefit people with dementia or cognitive impairment as they have been routinely excluded from participation in many of these studies.

The difficulties experienced by caregivers and dental professionals when providing oral health care to people with dementia are well documented (12). As a result, tailored behavioural and communication strategies are needed for this patient group. To date, however, evidence regarding the effectiveness of such interventions is lacking. Previous reviews have almost exclusively focused on the residential care setting and have not been limited to cognitively impaired populations. In this study, we reviewed the application and effectiveness of different interventions on the oral health of older people with dementia or cognitive impairment.

Methods

Search strategy

The Medline database was searched using a combination of the following medical subject headings (MeSH terms): dementia; “activities of daily living”; nursing staff; caregivers; hospitals; independent-living; long-term care; oral health; oral hygiene; tooth disease; dental caries; dental plaque; xerostomia; polypharmacy AND/OR the key terms listed in Appendix 1 in the title or abstract. Embase, CINAHL, Scopus and Cochrane databases were then searched using variations of this strategy based on database characteristics. All searches were conducted in January and February 2015.

Appendix 1.

MEDLINE SEARCH 06/02/2015

1. Dementia/

2. Alzheimer$.ab,ti

3. Cogniti$.ab,ti

4. (Cognitive adj1 impairment).ab,ti

5. “activities of daily living”/

6. (nursing adj1 home$).ab,ti

7. (care adj1 home$).ab,ti

8. Aged care.ab,ti

9. Nursing staff/

10. Nurse$.ab,ti

11. Long-term care

12. Caregivers/

13. Carer$.ab,ti

14. Hospitals/

15. (residential adj2 (care or facilit$3)).ab,ti

16. Independent living/

17. Community-dwelling.ab,ti

18. 1 or 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17

19. (editorial or letter).pt

20. 18 not 19

21. Limit 20 to (humans and “all adult (19 plus years)”)

22. Oral health/

23. Oral hygiene/

24. Exp tooth disease/ or dental caries/ or dental plaque/

25. (saliva or salivary gland or salivary gland dysfunction).ab,ti

26. Xerostomia/

27. Dry mouth. ab,ti

28. aspiration pneumonia.ab,ti

29. polypharmacy/

30. 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29

31. Limit 30 to (humans and “all adult (19 plus years)”)

32. 31 not 19

33. 21 and 32

34. Limit 33 to English language

35. Remove duplicates

Procedure

After excluding duplicates, all titles and abstracts were screened for relevance according to review eligibility criteria. To ensure reliability, a second reviewer screened 500 randomly selected titles and abstracts from the Medline database. There was complete agreement between reviewers. Full-text versions of articles were obtained and screened for inclusion. Quality assessment and data extraction were performed and checked by a second reviewer for consistency. There was high agreement between the two reviewers with regards to quality ratings assigned for study components (κ=0.866, p<0.001). Disagreements were resolved through discussion and consultation with a third reviewer. Back-references from all included articles were reviewed to detect potentially relevant papers not identified through the database search.

Inclusion criteria

We included interventions and implementation strategies designed to improve or maintain the oral health of older people (mean age >60 years) with dementia or cognitive impairment. Studies in which >50% of participants had dementia and/or cognitive impairment were also included as interventions found to be effective in these mixed populations may also have relevance for people with dementia and cognitive impairment.

The outcomes should be direct or proxy measures of oral health at the individual level (such as plaque, debris, gingival health or oral health-related quality of life).

Studies of any design that included an outcome comparison with either a randomly or non-randomly allocated comparison group or with baseline measures were included. Other criteria were language (published in English up until the date of the search) and sample size (≥5).

Data extraction and quality assessment

Data extracted from each study included bibliographic information and study source, study design, participant demographic characteristics, institution characteristics (if relevant), intervention description, outcomes, methods of evaluation and key results.

Study quality was assessed using the tool developed by the Effective Public Health Practice Project (13). This reliable, validated quality assessment (14) was developed specifically for systematic reviews of public health and health promotion interventions and is recommended by the Cochrane Handbook for systematic reviews of interventions (15).

The 21-item tool contains 6 sections assessing selection bias, study design, confounding, blinding, data collection methods and withdrawals/drop-outs. Each section is allocated a rating of ‘strong', ‘moderate' or ‘weak' based on pre-defined criteria. Each study is also ultimately assigned a global quality rating. For a study to be assigned a strong global quality rating, it must have no weak ratings and a minimum of four strong ratings across the six sections. A study is categorised as weak if it has two or more weak ratings.

Two other methodological considerations are included in the assessment tool: intervention integrity and appropriateness of quantitative analysis. While they do not contribute to the global quality score, they were reported for each study and used to inform the interpretation of results.

Results

The searches returned 11952 titles and abstracts. After removing duplicates and studies that did not meet the selection criteria, 115 abstracts remained. We were not able to obtain full papers for four studies (16, 17, 18, 19), two of which were published prior to 1990, so 111 papers were screened for inclusion. Seventeen studies met inclusion criteria and one additional eligible study was identified based on the included reference lists. This review finally included 18 papers (Figure 1).

Figure 1.

Flow diagram of the search and selection process

The studies were conducted in the US (20, 21, 22, 23, 24, 25), Japan (26, 27, 28, 29), Norway (30, 31), Belgium (32, 33), the Netherlands (34) Sweden (35), Switzerland (36) and Denmark (37). All studies were published between 1996 and 2014.

Due to significant heterogeneity among studies with respect to populations, outcomes and interventions, statistical methods were not used to pool results. Results are instead presented as a narrative review.

Table 1 summarises the basic characteristics of studies included for review. Eight studies employed an uncontrolled before and after design to assess intervention effectiveness (21, 22, 23, 24, 25, 28, 31, 35). The other 10 studies all included at least one comparison group in a variety of study designs; four were RCTs (27, 29, 30, 36), three were cluster RCTs (32, 33, 34), one was a quasi RCT (37), one was a controlled cross-over design (26) and one study employed a controlled before and after design (20). Six of the studies were described as pilot studies (20, 21, 22, 23, 24, 25).

Table 1.

Summary of intervention characteristics and results

Author (year) Study location; setting Design (study groups); intervention length Total n (% female); % with dementia or cognitive impairment (diagnostic criteria used)	Study groups and Intervention description	Outcomes (measurement): - Results	Quality rating*
Yoneyama et al. (1996) Japan; nursing home Controlled cross-over study (2 groups); 3 phases of 6 months each n=46 (67% f); 58.7% with ‘dementia or borderline dementia’ (Japanese MMSE)	Intervention group (2 groups; A [n=21] B [n=25]): - Phase 1 : Both groups received usual care (bedding, clothing and diaper changes as needed and sponge bath lx/week). - Phase 2: The A group received oral care from a dental hygienist lx/day and mouths cleansed by nurse after each meal by gargling with 1 % povidone iodine. The B group received usual care. - Phase 3: The B group received the oral hygiene intervention. A group received usual care.	Opportunistic infection (respiratory infection, febrile days, deaths from pneumonia): - no significant differences between groups in any phase, on any measures	+
Pyle et al. (1998) Midwestern United States; long-term care facility Controlled before and after study (2 groups); 6 weeks, with follow-up at 3 months n=23 (68% f); 73% had “significant mental impairments” (diagnostic criteria not specified)	Intervention group (n=12): AINs responsible for the care of intervention residents received 6 weeks of oral health education and training (1hr/week). Education included management of behavioural difficulties and instruction in preventative oral health regimens including use of sensonic toothbrushes. Individual oral care plans were developed for residents by research staff and implemented by AINs. Control group (n=11): Usual care (not described) provided by AINs who had not received the education intervention.	Dental hygiene (plaque index); gingival health (gingival index): - improvement in plaque and gingival scores over time was significantly greater in the intervention group compared to the control group	+
Connell et al. (2002) Virginia, United States; nursing home Case-series (5 individuals); 5 days n=5 (20% f); 100% with dementia (chart notation or MMSE≤24)	Intervention group (n=5): Following observation by a trained clinical research nurse, revised oral care plans were designed for each resident with the goal of removing barriers to greater independence and optimizing staff assistance with oral care. Care plan elements included modification to the physical environment (e.g. change sink height) and to staff practices in oral care provision (e.g. closing the door during oral care). Staff members attended a 30-minute in-service about oral care provision, supplemented by reminders posted in resident’s rooms.	Dental hygiene (plaque index): - all participants experienced reductions in oral plaque during the intervention Independence in oral care (nurse observation) [Secondary outcome]: - 2 residents were more independent post-intervention, 2 were more independent in at least 1/3 of tasks and 1 resident was less independent
Yoneyama et al. (2002) Japan; 11 nursing homes RCT (2 groups); 2 year n=366 (80% f); 68% with MMSE scores <21 (Japanese MMSE)	Intervention group (n=184): Nurses or carers brushed residents’ teeth with a toothbrush for approx. 5 minutes after every meal. If brushing was not efficient, the oropharynx was also scrubbed with an applicator and 1% povidone iodine. Control group (n=182): Usual care. Some residents brushed their own teeth but none requested help from caregivers.	Opportunistic infection (cases of pneumonia, febrile days, death from pneumonia); dental hygiene (debris index): - Residents in the oral care group experienced reduced risk of pneumonia and death from pneumonia relative to controls - no significant differences in febrile days - oral care significantly improved scores on the debris index compared to control (post-hoc) ADLs, cognitive function (MMSE) [Secondary outcomes] - decline in MMSE scores from baseline to 24 months was significantly less rapic in the oral care group compared to controls
Bellomo et al. (2005) Switzerland; Long-term care facility RCT (2 groups); 3 months n=61 (72% f); approximately 75% of residents included in analysis had moderate to severe cognitive impairment (MMSE scores ≤26)	Intervention group (n=29): Residents assessed as being Independent (n=16) received occupational therapy instruction on manual tooth and denture brushing. Those assessed as needing assistance (n=13) received the same instruction combined with monitoring (including guidance and gesture education during oral care) and re-education if necessary. Control group (n=30): After the baseline measures, independent controls received no intervention (n=15). Controls requiring assistance received weekly ‘placebo’ occupational therapy (e.g. manicure) with the same therapist (n=15).	Independence in oral care (therapist observation), Dental hygiene (plaque index) and denture hygiene (denture plaque index): - After 3 months denture brushing was performed more independently - Both intervention and placebo occupational therapy led to improvements in plaque and denture plaque but there was no significant difference between groups - Residents in the ‘ assisted intervention’ subgroup and those with a confirmed diagnosis of dementia had the greatest reductions in dental and denture plaque (post-hoc analyses)	++
Beck et al. (2008) Denmark; nursing home Quasi-RCT (2 groups); 11 weeks n=121 (73% f); approximately 75% with mild to severe cognitive impairment (MDS: CPS ≥2)	Intervention group (n=62): In addition to routine care, residents received nutritional supplementation, group exercise 2x/week and oral care 1-2x/week. Oral care was performed by a dental hygienist and consisted of manual brushing of teeth and/or dentures, palatal and mandibular mucosa and tongue dorsum. Interdental brushes were used for flossing. Control group (n=59): Routine nutritional care, physical therapy and oral care (not described)	Dental hygiene (MDS: plaque item): - 67% of oral care that was planned was completed but the prevalence of residents with visible plaque did not change (>30% at baseline)	+
Samson et al. (2009) Norway; nursing home Before and after study (2 groups; matched at 3 months, unmatched at 6 years); 3 months, with a 6 year follow-up n=88 (78% f at 6 year follow-up); 59% with manifest dementia (at 6 year follow-up)	Intervention group (n=88): Nursing staff received theoretical and practical education in oral care. Picture based procedure cards were developed for each resident and hung on their bathroom wall. Oral care appliances including electric toothbrushes were provided to each resident. New routines were implemented on the ward including appointment of an ‘ oral care contact’ person. Regular follow-ups and screening of residents’ oral hygiene were conducted by a dental hygienist and feedback provided to nursing staff.	Dental hygiene and mucosal condition (MPS: plaque score +mucosal score): - The percentage of residents with an acceptable MPS was significantly greater at 3 months and at 6 years compared to baseline	+
Kullberg et al. (2010) Sweden; nursing home Before and after study (1 groups); 3 weeks n=43 (71% f); 100% with dementia (no formal diagnosis but setting was a dementia care centre)	Intervention group (n=43): Staff who had received dental hygiene education 1.5 years previously received a second education session. The staff received individual instructions relating to each resident’s oral care needs as well as hands-on training in oral care. All residents with their own teeth were supplied with an electric toothbrush for use 2x/day and Chlorhexidine gluconate 1% gel for use 2 x/day for 1 week/month.	Dental hygiene (plaque score); gingival health (gingival bleeding score): significant reductions in plaque and gingival bleeding scores were observed 3 weeks after delivery of the education session	+
Jablonski et al. (2011) Pennsylvania, United States; nursing home Before and after study (1 group; pilot); 14 days n=7 (86% f); 100% with dementia (chart notation and medical records)	Intervention group (n=7): Residents were provided with mouth care based on the Managing Oral Hygiene Using Threat Reduction (MOUTh) intervention. The intervention combined evidence-based best mouth care practices for elderly people with strategies to reduce threat perception during the provision of oral care. Examples of threat reduction strategies provided were replacing one caregiver with another (rescuing) and smiling when interacting with the resident. Examples of best practice oral care included using interdental brushes between teeth and warm water for rinsing.	Care resistant behaviour [CRB] (revised restiveness to care scale) -a clinically but not statistically significant decrease was observed in the mean CRB rate from baseline to the observation period Oral health (Oral Health Assessment Tool [OHAT]) [secondary outcome] -Compared to baseline, mean OHAT scores improved significantly after 7 days and after 14 days of intervention	++
De Visschere et al. (2011) Flanders, Belgium; 14 nursing homes Cluster RCT (3 groups); 5 years, with collection points at 2 and 5 years n=1393 [but analysed sample much smaller](76% f); 70-75% with cognitive impairment (Katz’s index of independence in daily living)	Intervention group (n=211): Following an introductory information session for institution directors, RNs were appointed as oral health coordinators (OHC) and completed a half-day theoretical and practical training in oral care. A new procedure was implemented whereby all new arrivals to intervention homes received an oral health assessment by OHC, who then developed individualised care plans based on needs and level of dependency. Care plans were integrated into daily care and implemented by all care staff. Control group 1 (residents in intervention homes not receiving active intervention; n=511): usual care (not described) Control group 2 (Residents in control homes; n=671): usual care (not described)	Dental hygiene (plaque index); denture hygiene (denture plaque index): - The intervention was not a significant predictor of dental or denture plaque levels in a mixed methods analysis	++
De Visschere et al. (2012) † Flanders, Belgium; 12 nursing homes Cluster RCT (2 groups); 6 months n=373 (73% f); 86.5% with cognitive impairment (MMSE<26)	Intervention group (n=187): Supervised implementation of the Dutch “Oral health care (residential) care homes for elderly people” guideline. - A project supervisor, a physician and at least 2 ward oral health care organizers (WOOs) appointed at each facility. WOOs completed an informative oral presentation (1. 5hrs), lecture (2hrs) and practical education(3hrs) then provided theoretical and executive education session at each ward for nurses and aides (1. 5hrs). Daily oral care provided to residents in line with the guideline (not described) and investigators conducted monitoring visits at each facility every 6 weeks for listing and resolving implementation issues Control group (n=186): Oral care according to the unsupervised implementation of the “Oral health care (residential) care homes for elderly people” guideline (not described)	Dental hygiene (dental plaque index); denture hygiene (denture plaque score); tongue hygiene (tongue coating index): - A small but statistically significant beneficial effect of the intervention was observed for denture plaque at 6 month follow-up (after adjustment for baseline value and the random effect of institution). However, this effect was no longer statistically significant at the 5% level when random institution effect was included in a mixed-model linear regression. - No significant intervention effects were observed for dental plaque or tongue plaque.	++
van der Putten et al. (2013) † Netherlands; 12 nursing homes Cluster RCT (2 groups); 6 months n=342 (67% f); 53% of sample with primary dementia diagnosis; 74% of residents on psychogeriatric wards (n=187) had MMSE scores ≤24	Intervention group (n=177): Supervised implementation of the Dutch “Oral health care (residential) care homes for elderly people” guideline. - A project supervisor, a physician and at least 2 ward oral health care organizers (WOOs) appointed at each facility. WOOs completed an informative oral presentation (1.5hrs), lecture (2hrs) and practical education (1hr) then provided theoretical and executive education session at each ward for nurses and aides (1. 5hrs). Daily oral care provided to residents in line with the guideline (not described) and investigators conducted monitoring visits at each facility every 6 weeks for listing and resolving implementation issues Control group (n=165): Oral care according to the unsupervised implementation of the “Oral health care (residential) care homes for elderly people” guideline (not described)	Dental hygiene (dental plaque index); dental hygiene (denture plaque index): - At 6 months, mean dental and denture plaque levels were significantly better in the intervention compared to control groups - The multilevel mixed methods analysis conducted with plaque scores at 6 months as outcomes showed that the reduction by intervention was significant for denture but not dental plaque	++
Tashiro et al. (2012) Japan; nursing home Repeated before and after study (1 group; 3 interventions); 3 x 5 days of active intervention n=12 (75% f); 67% with senile dementia (medical records and staff interviews)	Intervention group (n=12): All 12 participants were administered 3 different oral care protocols for 5 consecutive days each, 3 weeks apart. Oral care was performed between lunch and dinner. Oral care protocols were (in the order received): 1. Manual tooth brushing for approximately 5 minutes 2. Tongue coat removal with a sponge brush 3. Wiping oral mucosa with a sponge brush soaked in 0.0002% chlorhexidine gluconate	Dental hygiene (plaque index); gingival health (gingival index); periodontal health (community periodontal index; bacteriological tests), tongue coating condition (according to accepted classification), dryness of mouth (saliva wetness test paper), oral malodour (observation by 2 raters); opportunistic infection (oral pathogens in unstimulated saliva) - Tooth brushing improved oral malodour as well as scores on plaque and gingival indices - Tongue coat removal improved tongue condition and oral malodour - Wiping oral mucosa with chlorhexidine decreased opportunistic infections in the pharynx region	+
Mentes et al. (2012) Los Angeles, United States; nursing home Before and after study (1 group; pilot); 21 days n=8 (63% f); not reported but the total CPS mean score was 2.38 (between mild and moderate impairment)	Intervention group (n=8): Residents were offered lollipops twice a day containing 2 mg of liquorice extract, hydrogenated starch hydrolysate, citric acid, colouring and flavouring agents, and sugar-free sweetener.	Microflora of the oral cavity (Streptococcus Mutans bacteria count) - Residents who consumed more lollipops over the study period tended to have a lower S. Mutans load Dental hygiene (debris and calculus score) - Debris and calculus scores showed no significant change over the 21 days	+
Sloane et al. (2013) North Carolina, United States; 3 nursing homes Before and after study (1 group); 8 weeks, with a 6 month follow-up at one site n=97 (75%f); 85% with mild to severe dementia (≥2 on MDS: CPS)	Intervention group (n=97): AINs appointed ‘oral care aides’ and provided with training in evidenced based mouth care and person centred behavioural care by a dentist and psychologist. Training and supervision was provided daily for 2 weeks and then decreased to a few hours per week. AINS were instructed to implement oral care protocols a minimum of 5 days/week. - Protocol for residents with natural teeth: teeth brushed and gingival tissues cleaned using 0.12% chlorhexidine rinse*; flossing with interdental brush dipped in 0.12% chlorhexidine; 1.1% sodium fluoride toothpaste applied to teeth and instructed not to eat for 30mins - Protocol for full or partial dentures: brush with toothbrush under running water; gingival tissues cleaned with 0.12% chlorhexidine* *after 6 weeks, chlorhexidine replaced with water, oral rinse or toothpaste	Dental hygiene (plaque index); gingival health (gingival index; MDS item-presence of bleeding gums); denure hygiene (denture plaque index); nutrition (nurse-reported food intake and problems with swallowing, body weight) - After 8 weeks scores improved significantly on the plaque, gingival and denture plaque indices but not on the MDS gingival inflammation item or reported food intake - Improved plaque, denture plaque and gingival indices were retained at 6 months in the one facility monitored (n=21). At 6 months, the MDS item was also improved compared to baseline.	+
Morino et al. (2014) Shizuoka, Japan; nursing home RCT (2 groups); 1 month of active intervention, with 3 and 5 month follow-ups n=34 (83% f); “most of the participants had cognitive disorders” (diagnostic criteria not specified)	Intervention group (n= 17): In addition to routine oral care (self-care or assisted by nursing staff), residents received professional oral health care (POHC) once a week after breakfast. POHC was performed by two dental hygienists and consisted of manual tooth brushing and flossing with interdental brushes. Dentures were cleaned with an ultrasonic cleaning apparatus, a toothbrush and denture tablets. Control group (n=17): Routine oral health routines were followed (self-care or assisted by nursing staff)	Microflora in the oral cavity (number or % of whole bacteria, Streptococcus, Fusobacterium and Prevotella); presence of opportunistic infection; dental hygiene (dental plaque index) and oral moisture - After 1 month residents in the intervention group had greater odds of experiencing an improvement in plaque scores compared to control residents (OR: 9.33 [95% CI = 1.74-75.66], p<.01) and greater odds of experiencing an increase in Streptococcus species in saliva compared to control residents (OR: 9.33 [95% CI = 1.29-193.09], p<.05). - All other effects were non-significant	++
Fjeld et al. (2014) Norway; 9 nursing homes RCT (2 groups); 2 months n=180 (75% f); 51.4% with moderate or serious cognitive impairment (physician diagnosis)	Intervention group (n=86): Residents were instructed to use an Oral-B Professional Care electronic oscillating toothbrush for oral care (ET). Residents and nurses were provided with individual instruction (written and picture cards) on tooth brush use by a dentist. Control group (n=94): Control residents were instructed to use a manual toothbrush (MT) as usual.	Dental hygiene (debris index from the Oral Hygiene Index-Simplified [OHI-S]); mucosal condition (mucosal plaque score) - No intervention effect was found for ET compared to MT for either mucosal condition or dental hygiene. However, both electric and manual tooth brushing led to an improvement in dental hygiene over the study period.	+++

* WEAK (+), MODERATE (++), STRONG (+++) methodological quality based on the Quality Assessment Tool for Quantitative Studies (12); †these studies were part of a large multi-centre study conducted concurrently in the Netherlands and Belgium Results are reported and published separately for the two locations; MDS: minimum dataset; AIN: assistant in nursing; MMSE: mini-mental state exam

Although no inclusion criteria were specified relating to study setting, studies were all set in nursing homes (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 35, 36, 37) or long-term care facilities (20, 34). The number of participants was highly variable, ranging from 5 to 373. While one study reported baseline characteristics of 1393 included participants, oral health outcomes were not reported for the whole sample (32).

A diagnosis of dementia was necessary for inclusion in only two studies (21, 38). Among other studies, the percentage of participants with dementia or cognitive impairment ranged from 51% (30) to 100% (35). A range of measures were used to assess cognitive impairment and diagnose dementia. Six studies used the English or Japanese versions of the Minimental State Exam (MMSE) (31, 26, 37, 33, 34, 36), three used the Cognitive Performance Scale (CPS) of the Minimum Data Set (MDS) (24, 25, 37), two used chart notation and medical records (28, 38) one relied on physician diagnosis of dementia (16) and one inferred cognitive impairment from scores on the Katz' index of Independence of Activities of Daily Living (32). While the remaining five studies discussed the cognitive status of included participants (e.g. “most of the participants had cognitive disorders” (29)), the diagnostic criteria used to draw these conclusions were not described (20, 22, 29, 31, 35).

Quality of included studies

Quality assessment resulted in 10 weak ratings, six moderate ratings and two strong ratings (Table 2). The most common methodological limitation was lack of blinding, and in 10 studies no attempt was made to blind either the participants or outcome assessors to allocation status. Management of putative confounders was also poorly performed or reported, with 10 studies failing to describe them at all, or controlling for fewer than 60% of relevant confounders. A number of studies were uncontrolled pre-post studies. Another limitation was the absence of a sample size justification; only two studies reported this information (33, 34). Intervention integrity was considered to some extent in many studies (21, 22, 23, 24, 25, 30, 33, 34, 35, 36, 37); however, not all authors reported these findings and intervention integrity was not always discussed in relation to results.

Table 2.

Summary of study quality

Study	Selection bias	Study design	Confounding	Blinding	Data collection methods	Withdrawal and drop outs	Global rating	Notes
Yoneyama et al. (1996)	WEAK	MODERATE	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity not measured or discussed
Pyle et al. (1998)	STRONG	STRONG	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity measured but not reported - no control for confounders - no reporting of raw scores
Connell et al. (2002)	WEAK	WEAK	WEAK	WEAK	STRONG	STRONG	WEAK	- some aspects of intervention fidelity considered but not discussed - no follow up beyond 5 days (pilot)
Yonehama et al. (2002)	MODERATE	STRONG	STRONG	MODERATE	STRONG	STRONG	STRONG	- intervention fidelity not measured or discussed
Bellomo et al. (2005)	WEAK	STRONG	STRONG	MODERATE	STRONG	STRONG	MODERATE	- intervention fidelity measured and discussed - results not clearly reported
Beck et al. (2008)	WEAK	STRONG	WEAK	MODERATE	STRONG	STRONG	WEAK	- intervention fidelity measured; only 60-79% received oral health as prescribed - no statistical results reported for oral health outcomes
Samson et al. (2009)	MODERATE	WEAK	WEAK	WEAK	STRONG	N/A	WEAK	- intervention fidelity not measured or discussed - statistical analysis (between groups ANOVA) not appropriate as two of the included groups appear to be pre and post measures from the same subjects
Quagliarello et al. (2009)	WEAK	MODERATE	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity measured; 80-100% of planned oral care was performed - between group comparisons made but confounders not considered
Kullberg et al. (2010)	MODERATE	MODERATE	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity measured and discussed - no follow-up beyond 3 weeks (pilot)
Jablonski et al. (2011)	MODERATE	MODERATE	WEAK	MODERATE	STRONG	STRONG	MODERATE	- intervention fidelity measured observationally and discussed - no follow up beyond 14 days (pilot)
De Visschere et al. (2011)	MODERATE	STRONG	STRONG	MODERATE	STRONG	MODERATE	MODERATE	- intervention fidelity not measured or discussed
De Visschere et al. (2012)	MODERATE	STRONG	STRONG	MODERATE	MODERATE	STRONG	MODERATE	- intervention fidelity measured (in process evaluation) but not reported in this publication
Van der Putten et al. (2012)	MODERATE	STRONG	STRONG	MODERATE	STRONG	MODERATE	MODERATE	- intervention fidelity measured (reported in separate process evaluation) but discussed as a potential limitation
Tashiro et al. (2012)	WEAK	MODERATE	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity not measured or discussed - statistics inappropriate and results not clearly reported
Mentes et al. (2012)	WEAK	MODERATE	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity measured; 60-79% of intervention was received as prescribed
Sloane et al. (2013)	MODERATE	MODERATE	WEAK	WEAK	STRONG	STRONG	WEAK	- intervention fidelity measured but not reported. However, ‘dose’ was added into the mixed model to assess its contribution to outcomes
Morino et al. (2014)	WEAK	STRONG	STRONG	MODERATE	STRONG	STRONG	MODERATE	- intervention fidelity not measured or discussed - possible that some control participants received unintended intervention (carry-over effects)
Fjeld et al. (2014)	MODERATE	STRONG	STRONG	MODERATE	STRONG	STRONG	STRONG	- intervention fidelity not measured but considered as potential limitation

Outcomes

Primary outcomes included a number of oral health measures, with most studies investigating more than one measure (Table 3). Dental hygiene, the most frequently reported outcome, was measured in 16 studies and was indicated by levels of dental plaque (20, 21, 22, 25, 28, 29, 31, 32, 33, 34, 35, 36, 37), dental debris (24, 27, 30) or calculus (24). Five studies also measured denture hygiene by assessing denture plaque quantity (25, 32, 33, 34, 36). However, there was inconsistency in the methods used across studies. The profession of the person performing the measure varied and many studies reported procedural modifications including which teeth were used to calculate plaque levels.

Table 3.

Summary of studies reporting a significant improvement in primary outcomes for the intervention group relative to control group or baseline measures, by methodological quality assessment¹

	Number of studies showing significant improvement in outcome (total number of studies measuring outcome)
	STRONG (n=2)	MODERATE (n=6)	WEAK (n=10)
Dental hygiene	1 (2)	1 (5)	6 (9*)
Denture hygiene	-	1 (4)	1 (1)
Gingival health	-	-	4 (4)
Mucosal condition	0 (1)	-	1 (1)
Tongue hygiene	-	0 (1)	1 (1)
Presence of opportunistic infection	1 (1)	0 (1)	1 (2)
Microflora of the oral cavity	-	1 (1)	0 (1)
Swallowing and aspiration	-	-	1 (1)
Nutritional status	-	-	0 (1)
Care resistant behaviour	-	0 (1)	-
OHAT	-	1 (1)	-
Independence in oral care	-	1 (1)	0 (1)
Oral moisture	-	0 (1)	0 (1)
Oral malodour	-	-	1 (1)

¹ Based on the Quality Assessment Tool for Quantitative Studies; *one study did not statistically test improvements in dental plaque (1); outcome was not measured in any of the studies (-); OHAT: Oral health Assessment Tool.

Two studies measured mucosal condition by considering mucosal plaque (30, 31), and one assessed periodontal health using the community periodontal index (28). Gingival health was measured by the gingival index (20, 28), the bleeding gums item in the MDS (25) or gingival bleeding score (35). The microflora of the oral cavity was measured in two studies (24, 29); one study measured Streptococcus mutans load in unstimulated saliva and another measured Streptococcus, Fusobacterium and Prevotella levels. Tongue plaque was used as a measure of oral hygiene in two studies (28, 33). One study used the Oral Health Assessment Tool (OHAT) as a general measure of oral health (23). Swallowing and aspiration was assessed in one study by considering cough frequency during meals (22). Another study assessed nutritional status by measuring modifications to food intake and changes in body weight (25). Opportunistic infection was an outcome in three studies and was measured in a variety of ways; febrile days, cases of pneumonia and deaths from pneumonia were considered in one study (27), markers of respiratory infection including abnormalities of CRP, white blood cell count, alpha-2 globulins and pathogens in sputum were considered in another (26) and one study considered the presence of opportunistic pathogens in unstimulated saliva (29).

Interventions and effectiveness

The intervention characteristics of included studies were highly variable. To assist in the presentation and interpretation of results, as well as maintain consistency with earlier reviews (3), interventions were broadly grouped into four different categories: a) oral hygiene care strategies, b) behavioural and communication strategies, c) oral health care training and oral care provision by staff and carers and d) comprehensive oral health protocols. Some studies fit into more than one category but are presented here only once.

Preventive oral hygiene care strategies

Professional oral health care (POHC) and assisted brushing: Five studies investigated the effect of POHC on oral health outcomes (26, 27, 28, 29, 37). POHC was provided as a supplement to standard oral care in all but one study (28), in which POHC was the primary form of oral care. Study periods ranged from 5 days of active intervention (28) to 2 years (27). POHC was performed by both dentists and dental hygienists (26, 27, 28) or exclusively by dental hygienists (29, 37). The frequency of POHC ranged from once or twice a week (27, 29, 37) to once a day (26, 28).

The techniques and equipment used during POHC varied among studies. One study examined the effect of three different POHC protocols: oral cleaning using a manual toothbrush, tongue coat removal using a sponge brush and wiping the oral mucosa with a chlorhexidine solution (28). Two studies used manual toothbrushes to brush natural teeth and interdental brushes for flossing. In one of those studies, dentures were brushed the same way (37) but in the other a denture cleaning tablet and ultrasonic cleaning apparatus was used for dentures (29). The techniques and equipment used during POHC were not described in 2 studies (26, 27).

The results pertaining to POHC were mixed, with three studies reporting significant effects on at least one outcome measure (27, 28, 29). The only strongly rated study of POHC found significant improvements in number of patients with fever, number of patients with pneumonia and deaths from pneumonia following oral care (27). The authors also reported that the decline in MMSE scores from baseline to 24 months was significantly less rapid in the oral care group than in controls. In two other studies, POHC was found to improve oral malodour, dental plaque and gingival indices (28), decrease opportunistic infections (28) and improve microbiological parameters (29). Two studies investigating the effect of POHC found no significant improvement in plaque (37) or respiratory infection (26).

Electric toothbrush: One study compared the effect of an electric toothbrush on dental hygiene with manual tooth brushing (30). Residents were instructed to perform oral care twice a day including brushing natural teeth, interdental hygiene and cleaning of dentures. Residents and their carers were given oral and written (illustrated instructional cards) instructions on how to use the electric or manual toothbrush. Brushing with an electric toothbrush led to a significant reduction in plaque over the study period. However, the improvement was not significantly different from manual tooth brushing.

Liquorice root: One study examined the effect of a sugarfree herbal lollipop containing liquorice root on cariogenic bacteria levels in nursing home residents (24). Participating residents were offered lollipops twice a day. The study found no significant improvement in oral microflora over the study period, but did observe a non-significant differences in levels of Streptococcus mutans, with lower levels in residents who consumed more lollipops.

Occupational therapy: One study employed occupational therapy techniques to teach and supervise oral hygiene practices to residents in a long-term care facility (36). Based on their assessed level of independence with oral care, residents in the experimental group received either occupational therapy instruction on tooth and denture brushing alone (independent residents) or supplemented with weekly monitoring including guidance and gesture education during oral care (for residents needing assistance). The improvements in dental and denture hygiene experienced by the experimental group over the study period did not differ significantly to the control group. However, there was some evidence that residents in the experimental group who received weekly monitoring and gesture education improved more than residents receiving only initial occupational therapy instruction.

Behavioural and communication strategies

Two studies investigated the effect of behavioural and communication strategies. Jablonski and colleagues (38) piloted an intervention to reduce care-resistant behaviour during oral care in people with dementia. The intervention combined best practice mouth care with strategies designed to reduce the perception of threat during oral hygiene care provision. Mouth care was provided twice daily for two weeks by trained research staff and included brushing with a soft toothbrush, flossing with interdental brushes and mouthwash for rinsing. Threatreduction techniques were a collection of behavioural and communication strategies derived from nursing, neurobiological and dental research such as gesturing, establishing rapport and cueing and were tailored to individual residents through a process of trial and error. The intervention led to a significant improvement in oral health measured by the OHAT after 14 days and a reduction in care-resistant behaviour that approached significance (p=.06).

Sloane and colleagues (25) trained nursing assistants in evidence-based oral care techniques and person-centred behavioural care. Training was provided by a dentist and psychologist and consisted of both seminar and hands-on training in a peer-to-peer approach. Training and supervision were provided daily for two weeks, after which mouth care was provided a minimum of five days per week by nursing assistants. Mouth care included manual brushing of teeth with chlorhexidine rinse, swabbing of gingival tissue with chlorhexidine soaked swabs, flossing with interdental brushes dipped in chlorhexidine, application of sodium fluoride toothpaste to teeth and instruction not to eat for 30 minutes. Dentures were brushed with a toothbrush under running water and gingival tissues were cleaned with chlorhexidine soaked swabs. Person-centred behavioural strategies included approaching in a calm manner, using eye contact and gentle touch. After 8 weeks, the intervention led to significant improvements in dental plaque, denture plaque and gingival health.

Oral health care training and oral hygiene care provision by staff and carers

Four studies reported training programs and oral hygiene care provision, with training aimed at nurses or nursing assistants in residential aged care facilities. Two studies reported the profession of the educators (health educator (22) and dental hygienist (35)). The time spent on education varied considerably from one 30-minute session (21) to six weeks of hourly sessions (20, 22). One study reported a single in-service training session of unreported length (35).

The topics covered in the education programs included the rationale of the study (22), general information about ageing anatomy and the importance of oral care in the elderly (20, 21), barriers to good oral care in people with dementia (20) and management of behavioural disturbance during oral care (20). In all studies the education program included an overview or demonstration of oral hygiene techniques and steps for performing oral hygiene protocols (20, 21, 22, 35). Two studies also provided the opportunity for practical, hands-on training in oral care (20, 35) and one reported that caregivers were given the opportunity to ask questions about practical implementation (22). In one study, the education and training session was supplemented by a poster presentation available to all care home staff and reminders to perform oral care placed in residents' rooms (21).

In three studies, education preceded the implementation of specific oral hygiene protocols by staff and carers (20, 21, 22). In one study, residents with impaired oral hygiene were supported to use a manual toothbrush and perform a chlorhexidine rinse daily at two different frequencies. Residents with swallowing difficulty either were taught swallowing techniques by the trained nurses, received manual oral brushing or performed oral health care in an upright feeding position (22). In two other studies the trained nursing assistants were responsible for implementing individual care plans designed for residents by research staff (20) or a clinical nurse involved in the study (21). In one study the care plan was specifically designed based on occupational therapy principles to remove barriers to independence in oral care. The care plan included modifications to the physical environment and instructions to staff to make use of preserved abilities and overcome cognitive deficits (21).

All four studies reported significant results on primary outcome measures. Oral health care training and oral hygiene provision by staff and carers was shown to lead to reductions in plaque (20, 21, 22), improved gingival health (20, 35) and improved swallowing (22). The only study to measure secondary outcomes found no clear evidence of improvements in independence during oral care (21). All studies were judged to be of poor methodological quality.

Comprehensive protocols/complex interventions

Four studies evaluated the implementation of a comprehensive oral hygiene protocol in nursing homes (31, 32, 33, 34). In all studies, members of nursing staff were appointed to oral health coordinator/organizer (OHC) positions, responsible for protocol implementation in their wards. In one study, OHCs were also responsible for assessing the oral health of new arrivals to the nursing home and developing individualised oral care plans to be integrated into daily care (32).

OHCs were provided with a mixture of theoretical and practical training, including on how oral care should be performed and how often. The specifics of the oral care provided to participating residents were not described in any of the four studies. In one study, picture-based procedure cards were produced highlighting the individual care needs of residents in order to facilitate learning and implementation of the oral care protocol (31). The total duration of training received by OHCs was between half a day (32) and 6.5 hours (34). In all studies, OHCs were responsible for peer-to-peer training of other care staff. Two studies reported that OHCs were specifically trained to educate other care staff (33, 34). The same two studies were also the only studies to report the amount of time OHCs spent training other care staff (1.5 hours). Three studies reported that investigators or dental professionals performed monitoring visits to intervention sites to follow up on implementation (31, 33, 34).

The findings on comprehensive protocols were mixed. All but one study reported significant effects for at least one primary outcome. One study reported an increased percentage of residents with an acceptable level of oral health after three months and after six years, although this study was uncontrolled and suffered from a number of methodological limitations (31). The remaining three studies were all judged to be of moderate quality; two of these studies reported a significant effect of the intervention on denture plaque but not dental (25, 26) or tongue plaque (33), and one found that the intervention was not a significant predictor of either dental or denture plaque (32).

Discussion

Our most striking finding was the lack of high-quality evidence supporting the efficacious use of oral health interventions and implementation strategies for older people with dementia or cognitive impairment. Despite the high importance of oral health for this population, a comprehensive literature review identified only 18 relevant studies. Of these, only two interventions were identified that were designed specifically for and examined exclusively in people with dementia. Despite our broad inclusion criteria, no study was identified that had been conducted in community-dwelling adults with dementia or cognitive impairment. The reviewed studies varied considerably in terms of size, scope and focus. Many methodological weaknesses were identified with most studies judged to be of low or moderate methodological quality. Although two thirds of interventions included reported positive results for at least one primary oral health outcome, the specific outcomes and methods used to assess effectiveness varied and their overall quality were poor.

Both the heterogeneity of reviewed studies and the overall weakness in the quality of the study protocols limited our ability to identify the most effective strategies and intervention components for improving the oral health of people with dementia and cognitive impairment. Only one of the two high quality studies reported positive outcomes (27). Yoneyama and colleagues found evidence that POHC provided weekly as a supplement to routine oral health care reduced the risk of opportunistic infection in terms of pneumonia, death from pneumonia and febrile days. The other high-quality study showed that brushing teeth twice daily with either a manual or electric toothbrush improved dental hygiene, but that an electric toothbrush was no more effective than manual brushing (30). While these studies may provide preliminary evidence for the efficacy of these isolated interventions, it is not possible to draw clear conclusions based on the limited evidence available.

Strengths and limitations

This review outlines the current published evidence regarding oral health interventions for older people with dementia or cognitive impairment and provides direction for future research in this emerging field. This is the first review on this topic that has attempted to look only at studies conducted predominantly in cognitively impaired populations. We identified a large number of relevant studies through a comprehensive search strategy using a broad range of academic databases. We also attempted to increase validity by having two assessors and using a validated quality assessment tool.

Some limitations warrant discussion. Firstly, the search was limited to studies available electronically (and in full) and those that had been published in English, which may have led to publication bias. All included studies were conducted in nursing homes or long-term care facilities and in developed countries, which limits the generalisability of findings. Further, the majority of studies were not evaluated exclusively in those with dementia and cognitive impairment. It is possible that the findings may be more relevant to the general population of aged care residents.

Perspectives

The review included all studies regardless of methodological quality, which limited the conclusions that could be drawn about intervention efficacy. However, this approach did enable us to identify not only that study quality tends to be poor in this research area, but also to determine the specific methodological limitations that must be rectified in future research. Recommendations for research design are summarised in Table 4.

Table 4.

Summary of key recommendations for future research

Domain	Recommendation
Sample	- Specify recruitment method, setting and population (e.g. cognitive impaired, dementia, residential care, community-dwelling) - Use validated measures of cognitive status - Provide sufficient demographic and cognitive information - Perform power analysis to determine sample size
Design	- Pilot studies, when successful, to be followed up by larger scale effectiveness study (e.g. RCT) - Consider setting, barriers to care and target group in intervention design - Evaluation approach should inform intervention design - Follow-up for sufficient time to establish sustainability of intervention - Include relevant control group
Intervention	- Replicable, specific - Explicit description of study activities and processes - Feasible within context
Outcomes	- Consider broader range of and better quality outcomes
	- Choose outcomes that are specific, validated and reliable - Consider subject’s perspective in choice of outcomes (e.g. QoL may be more relevant than levels of dental plaque) - Consider proximal, intermediate and distal outcomes
Analysis and reporting	- Report effect size - Discuss clinical significance - Measure and discuss intervention integrity (beyond compliance) - Mixed methods may be appropriate in some cases (e.g. complex interventions, pilot studies) - Identify and control for relevant confounders, where possible

Many studies made only limited attempts to reduce bias resulting from insufficient randomisation, lack of an adequate control group, insufficient blinding of assessors and failure to control for relevant confounders. Only two studies justified their sample size, which made it impossible to determine whether unpromising results were a reflection of the intervention itself or due to a lack of statistical power. Of those that did report statistically significant effects, most did not discuss clinical significance and none reported an effect size. This is an important omission because improvements in dental plaque or gingival health are more meaningful if they also have practical or applied value to older people with dementia and their caregivers.

There was often little or no effort to evaluate the impact of interventions on the oral health of people with dementia over a sustained period. While there are challenges associated with longitudinal research in older adults, including attrition and high caregiver turnovers (39), methods at the design and analysis stages can be used to minimise the risk of bias associated with attrition. In addition to the recommendations listed in Table 4, these may include a combination of different retention strategies and incentives, as well as appropriate strategies at the analysis stage, depending on the randomness of missing data.

Another problem was the narrow range and poor quality of outcome measures used. Poor oral health is associated with a range of significant negative outcomes for older people, including poor general health, self-esteem and quality of life (40). However, the studies almost exclusively focused on proximal clinical oral health outcomes such as levels of dental or denture plaque despite validated measures to assess more distal outcomes such as wellbeing and oral health-related quality of life (OHRQoL) in older adults being available; none of the reviewed studies measured these constructs. These psychosocial measures are likely to be more salient to older people with dementia and their caregivers than clinical oral health outcomes and should be incorporated into future evaluations (11).

The complexity of the interventions further limited our ability to identify those intervention components that may be most effective in improving the oral health of people with dementia and cognitive impairment. The majority of the studies included more than one intervention component and, at least implicitly, targeted more than one barrier to care. In theory, multifaceted interventions that target multiple factors or barriers to care should be more effective than single component studies (41). Unless interventions clearly specify all components and the intermediate links between intervention activities and expected outcomes, they often fail to provide useful information (42). Most reviewed studies provided a detailed description of intervention components but few described a theoretical model specifying key elements and how they were expected to achieve intended outcomes. In educational interventions, for example, the intermediate links between educational components and oral health outcomes, such as improved staff attitudes, knowledge or skill, were neither routinely specified nor evaluated. Consequently, it was difficult to identify which intervention component/s conferred the most benefit, or contributed most to its failure.

While most reviewed studies did make some attempt to measure intervention integrity, these attempts tended to focus on compliance at the expense of other aspects of integrity, and integrity was not always considered in relation to outcomes. A singular focus on compliance is unlikely to be appropriate in oral health interventions where competence or quality of delivery is also likely to be an issue. Although some studies examined the quality of delivery of the prescribed oral care protocol, few interventions examined the quality of delivery of educational and training components, despite its potential relevance to the outcomes of the study. In studies that used a peer-to-peer teaching approach, for example, little emphasis was placed on the quantity and quality of knowledge transmitted, the time spent, or the number of staff eventually reached.

The success of oral care interventions in residential care settings depends as much on organisational structures and culture as on the provision of best practice treatment and resources (42). However, few studies discussed the role of institutional factors such as lines of authority, staff workloads and philosophical values in the ultimate success or failure of the studied interventions, making it difficult to determine whether the intervention may be effective in other settings.

Overall, we believe improved analysis and reporting is required, including discussion of effect size and clinical significance, and measurement of intervention integrity (beyond compliance). Mixed methods may be appropriate in some cases (e.g. complex interventions, pilot studies), and relevant confounders should be identified and controlled for, where possible.

Conclusions

Although there was evidence that some isolated interventions may have positive effects on oral health in the short term, it is not possible to make recommendations about the benefit of specific interventions for people with dementia or cognitive impairment based on the available evidence. Many reviewed studies conducted on this topic showed severe methodological weaknesses. The overall quality of evaluation must be improved and more emphasis needs to be placed on developing effective and sustainable strategies that address the needs of people with dementia.

Acknowledgements: Dr Katrin Seeher assisted with interpretation of statistical methods, Dr Peter Foltyn assisted with proof reading and commentary. Funding for the review was provided by the Dementia Collaborative Research Centre, UNSW, Australia.

Conflicts of Interest: Over the last three years, Henry Brodaty has been on advisory boards of or a consultant to Eli Lilly, Merck and Nutricia. His department has received payment to participate in drug trials for Alzheimer's disease by Merck, Sanofi, Servier, Eli Lilly and Tau Therapeutics. He has been recipient of grants for research by the National Health and Medical Research Council and Australian Department of Health and Ageing. Ms. Siegel has nothing to disclose. Ms. Cations has nothing to disclose. Dr. Wright has nothing to disclose. Dr. Naganathan has nothing to disclose. Dr. Deutsch has nothing to disclose. Dr. Brodaty reports other from Eli Lilly, Merck and Nutricia, grants from Merck, Sanofi, Servier, Eli Lilly and Tau Therapeutics, outside the submitted work.

Ethical standards: This literature search was a component of a clinical study of nurses assessment of oral health and preventive interventions. It was approved by the Sydney Local Health District Human Research Ethics Committee CRGH #CH62/6/2014-107.