Study design
The study is a cluster randomized controlled trial. The clusters, Dutch nursing homes, are randomly assigned to either the intervention condition (implementation of the STA OP!-protocol) or the control condition (general training without the step-wise approach). Measurements take place at baseline, after 3 months (at the end of the STA OP!-training period) and after 6 months.
The study is single blinded i.e. the researcher will know the condition, while the research assistants performing the measurements and the persons analysing and score the video registrations are blinded. They will also not be informed about the specific research questions and conditions.
Procedure
Recruitment
Institutions and nursing homes will be recruited within the 'University Network for Organisations of Elderly care' of the VU University Medical Center (UNO-VUmc). The aim of this collaboration network is to build greater knowledge about the best multidisciplinary care for vulnerable elderly. The network provides an infrastructure in which patient care/health care innovations, education and research can take place. Cooperation between institutions and between the institutions and the university will lead to more knowledge and better care. More than half of the eighteen member institutions will participate in this study.
Inclusion criteria of units
The nursing homes participating in this study will be randomized in an intervention or a control condition. Nursing homes participating in this study have to meet the following criteria: at least one psychogeriatric unit willing to participate and no major organizational changes or building activities are planned or performed in the study period. Facilities are invited to participate in the study, and after approval/inclusion randomized into control or intervention.
Randomization
A separate number will be assigned in order of registration to every institution, nursing home and/or participating unit. When an institution is participating with multiple affiliated nursing homes and/or units, we will investigate whether there are sources of possible contamination by professionals. Contamination is possible, if a professional (e.g. nurse, physician or psychologist) is operating in various nursing homes and/or units within the institution. If so, these different affiliated nursing homes and/or units will be randomized as one. In the event that an institution with multiple affiliated nursing homes and/or unit is participating in the study and there is no contamination by professionals, these locations and/or units will be randomized as separate participants. The randomization will take place by an independent researcher using the program "Random Allocation Software", made available by the EMGO+ Institute for Health and Care Research in Amsterdam.
Informed proxy consent will be obtained for all residents by family/caregivers. Residents will then be screened on dementia severity, behavioural problems and indications of pain.
Inclusion and exclusion criteria (residents) (Table 1)
Table 1
Instruments and the different assessment points
Selection criteria
| - GDS (Reisberg) | Pre selection | NHP | | |
| - MDS-CPS | Pre selection | NHP | | |
Primary Outcome
| - CMAI | Incl. | Nurse | Nurse | Nurse |
| - MDS-RAI pain scale | Incl. | Nurse | Nurse | Nurse |
| - NPI-NH | Incl. | Nurse | Nurse | Nurse |
| - PACSLAC-D | Incl. | Nurse & VIDEO | Nurse & VIDEO | Nurse |
Secondary Outcome
| - Qualidem | Incl. | Nurse | Nurse | Nurse |
| - Cornell & MDS-DRS | Incl. | Nurse | Nurse | Nurse |
| - Medication | Incl. | RA | RA | RA |
| - DS-DAT | Incl. | VIDEO | VIDEO | |
| - FACS | Incl. | VIDEO | VIDEO | |
Control variables/Demographics
| - (co-) Morbidity | Incl. | NHP | | |
| - Hospitalisation, fixation etc. | Incl. | RA | RA | RA |
| - Demographics; sex, age, background | Incl. | RA | | |
| - ADL (Katz) | Incl. | Nurse | | |
| - Apolipoprotein E4 | Incl. | RA | | |
Residents to be entered into the study meet the following (ABC)
pre-selection criteria:
A)
Moderate to severe cognitive impairment, according to the Global Detoriation Scale (GDS) [
29]. Those residents with a GDS score of 5, 6 or 7 will be eligible for this study.
B)
No chronic psychiatric diagnosis other than a dementia associated diagnosis.
The GDS score and also the absence or presence of a chronic psychiatric score (see criteria 1 and 2), will be assessed by the nursing home physician.
Proxy consent will be obtained for all residents meeting these two criteria (A & B). However, residents who are to be actively enrolled in the experimental or control condition will have to meet an additional
selection criterion.
C)
The criterion to enter the study protocol is the amount of clinically significant symptoms of pain and/or challenging behaviour. Inclusion criteria are: a score of at least 44 on the Dutch version of the Cohen-Mansfield Agitation Inventory (CMAI) [
30,
31]
or a score of at least 4 (frequency × severity) on items of the Neuropsychiatric Inventory - Nursing Home Version (NPI-NH) [
32,
33]
or an indication of clinically relevant pain (intensity × frequency ≥ 2) according to the Minimum Data Set of the Resident Assessment Instrument (MDS-RAI)-pain scale [
34] in measurement week 0 (baseline).
The following data will be gathered for all the included residents: Pain observation (Dutch version of the Pain Assessment Checklist for Seniors; PACSLAC-D) [
35‐
37] Depression (Cornell and MDS-RAI depression scale) [
38,
39] Quality of life (Qualidem) [
40‐
42], ADL (Katz) [
43], co-morbidity, demographics and Apolipoprotein Epsilon 4 (from Buccal mucosa swabs). In addition, 5-minutes video recordings will be made for a blinded scoring of pain and/or discomfort.
This inclusion procedure will take place on units in the intervention condition, as well on units in the control condition. These control units will be instructed to treat the included residents with 'standard care'.
'Clinical' inclusion of residents
Those residents who initially do not score above the threshold on the NPI-NH, MDS-RAI pain scale or the CMAI, but who have a change in behaviour after the instrument screening, can be included in the study (after contacting the researcher and after obtaining proxy informed consent). Newly admitted residents on the unit can, after a six week 'adjustment period', follow the same procedure. The resident will (again) be assessed with the same instruments as the regular residents included, after approval by the researcher. Final analysis will be performed with and without these additional groups.
Sample size and power analysis
One of the primary outcome measures in this study is the CMAI, a behavioural observation scale with 29 behavioural items. Each item may be scored between 1 and 7, depending on the frequency of existing symptoms. The range on the CMAI is 29-203. In Dutch nursing homes the median is 44 [
44]; there is no normal distribution, but the distribution of the change is more or less normally distributed. The standard deviation of difference scores (difference of 4) is 13. The expected difference is however larger (at least 10 points). To detect a 15% difference between the intervention (STA OP!-protocol) and control condition with an α = .05 and β = .80, 56 residents will be required. However, in cluster randomization, a design effect should be taken into account; clustering of items (consistency of CMAI items [ICC] among residents in a unit is .10) and also the effect of large differences in sample sizes per cluster enhances the design effect [
45].
In this study the design effect was estimated at 1.5; So, 1.5 × 56 = 84 residents (after 4 weeks) will be needed in total to detect a 15% difference with an α = .05 and β = .80. However, we expect a 50% dropout due to non-response and loss to follow-up, i.e. we will need 168 patients in total; 84 in the intervention- and 84 in the control group.
Intervention condition: implementation of STA OP!
A comprehensive training program for healthcare professionals will be organized to implement the STA OP!-protocol. Healthcare professionals participating in this study will be nursing home physicians, psychologists, occupational therapists and registered or certified nurses (CNA/RN). In most of the nursing homes the majority of the registered or certified nurses will have relatively low educational levels 1, 2 or 3, which in Dutch is called 'verzorgenden' or 'helpenden'. Nurses and physicians who only work night shifts will be excluded from this study. Only the CNA/RN (levels 3, 4 or 5) is authorized to start the protocol.
Nursing home staff from all experimental units will be trained in the stepwise working method of the STA OP!-protocol in five meetings lasting three hours each. The team members will be taught enhanced physical and affective assessment skills that target needs commonly found in people with advanced dementia. A collaborating training centre, with very experienced trainers, will provide the STA OP!-training. These trainers are advanced practice nurses or have other medical backgrounds and have specific professional expertise regarding dementia, pain and discomfort.
In the sessions it will be discussed how nursing home staffs can recognize symptoms related to pain and affective discomfort, and how they can communicate with each other about these symptoms. In the first meetings the initial two steps of the STA OP!-protocol will be discussed, while in the follow up meetings the focus is on the last three steps.
The steps of the STA OP!-protocol are outlined in the following section. Each step is systematically written down in the STA OP!-protocol, which serves as a template for nurses and nursing home physicians to use the stepwise approach in practice (see also Figure
1). How much time it takes to decide to proceed to the subsequent step, depends on the specific interventions chosen. However, in general when effects are lacking or when they are limited, this should not take longer than one week.
Other healthcare professionals can be consulted at any time and at each step in the process.
• Step 0: Basic care needs
The nurse assesses whether or not basic physical care needs are fulfilled. Basic care needs concern, for instance hunger, thirst, a need for glasses, hearing aids or toileting.
• Step 1: Physical needs assessment (nurse and nursing home physician)
In addition to a brief physical nursing assessment, the nurse will also fill out an observational pain instrument (PACSLAC-D). If these assessments are negative, a nursing home physician (or if available a nurse practitioner) performs a more comprehensive physical assessment to find other probable physical causes, such as inflammation, infection, acute illness or a chronic condition possibly responsible for the observed behaviour.
For those residents already using pain medication or psychotropic drugs and still have behavioural symptoms possibly related to pain or affective discomfort, the nursing home physician will also assess whether the medication given is in accordance with the guidelines of the World Health Organization (WHO) and Verenso (the Dutch association of nursing home physicians) (also see Step 4 and 5).
• Step 2: Affective needs assessment
By using a needs-oriented and tailored approach, the nurse assesses possible problems regarding environmental stress, a possible imbalance between sensory stimulating and sensory calming activities, or a lack in meaningful human interactions. The psychologist working in the nursing home can be consulted at this step.
• Step 3: Non-pharmacological comfort interventions
In this step non-pharmacological comfort interventions will be conducted and implemented, in line with the personal history of the resident. Examples of comfort interventions are soothing, supportive verbal communication, supportive touch, and sensory stimulation by music, nice smells or soft materials. A study by van Weert [
46] showed that these types of comfort interventions are effective in reducing discomfort in nursing home residents.
• Step 4: Trial of analgesics
In this step of the protocol the nursing home physician is advised to prescribe a trial of analgesics according to the validated pain ladder, developed by the World Health Organization [
47]. Specific guidelines for use in the elderly are given to each participating nursing home physician in a training session, and similarly for physicians working in control and intervention units.
• Step 5: Consultation of relevant other disciplines (e.g. psychiatrist) and/or trial of prescribed psychotropic drugs
In the protocol and training sessions, nursing home physicians are instructed to use the guidelines of the Dutch association of nursing home physicians (Verenso) for prescribing psychotropic medication [
48]. The Verenso-guidelines clearly describe how and when psychotropic drugs are beneficial for dementia patients. In general, it is believed that this medication is only indicated for specific symptoms and for a fixed period of time.
In the training session it is also discussed that in general, a subsequent protocol step is needed when targeted assessments are negative, or when the symptoms related to pain or affective discomfort have not been reduced sufficiently by the targeted interventions.
During the training, the nurses and nursing home physicians start using the STA OP!-protocol for the residents who have been assessed by the blinded research assistants at pre-test (in week 0) as having a CMAI, NPI-NH and/or MDS-RAI pain scale score above the threshold. It can be expected that the STA OP!-protocol will also be used in other residents showing behavioural symptoms of pain or affective discomfort (outside the measurement weeks): this is allowed and will be registered. To promote the use of the protocol in practice, a focus group will be formed within the units of the institution. The researcher performs site-visits ones a week on the experimental units and provides nurses and nursing home physicians feedback on their use of the STA OP!-protocol in particular residents and answers their questions regarding pain or affective discomfort.
Control condition
In the control condition there is also a training program for healthcare professionals/nurses focusing on challenging behaviour and pain, but without the stepwise approach.
The nursing home physicians (and nurse practitioners if available) in the experimental and control group receive, besides the training sessions focusing on challenging behaviour and pain (with or without a stepwise approach), both exactly the same training regarding the (evidence based) treatment of pain and challenging behaviour. The training in geriatric pain management focuses on appropriate short- and long-acting drugs for the treatment of acute and chronic pain, dose escalation, analgesic escalation and management of side effects. The training in geriatric behaviour treatment focuses on the appropriateness of the use of psychotropic medication for several indications and its side effects. In this additional training it is stressed that other interventions are often more appropriate. This training is given by an experienced nursing home physician (WA).
Next to these two training programs, there is one important change in the control condition from the standard care situation: the nurses and nursing home physicians are informed which residents of their units have a CMAI, NPI-NH or MDS-RAI pain scale score higher than threshold at pre-test (week 0), according to the assessments of the nurses and blinded research assistants.
Therapy compliance
In the proposed study, therapy compliance is taken into account by the fact that the research assistant/advanced practice nurse who visits the experimental units once a week will have ample and structured focus on the actual use of the protocol in practice. The site-visits will include assessment whether the protocol is adequately used, the patient assessments performed adequately, if planned non-pharmacological interventions are executed and if prescribed analgesics or psychotropic drugs are indeed administered and used. The daily logs from the units will provide additional information on the actual use of the protocol, as well as reports from interdisciplinary meetings.
Primary outcomes
Primary outcomes measures include symptoms of challenging behaviour and/or pain.
Challenging behaviour
Symptoms of challenging behaviour will be measured with the Dutch version of the
Cohen-Mansfield Agitation Inventory (CMAI) [
30,
31,
49,
50], and with the Dutch version of the Neuropsychiatric Inventory-Nursing Home Version (NPI-NH) [
51‐
53].
The CMAI was developed to assess agitation in nursing home patients [
30,
31]. It is more specific than general-purpose behaviour rating scales, which usually include self-care activities, cognition, and mood items. The CMAI uses a 7-point scale to assess the frequency (ranging from 'never' to 'several times an hour') of 29 behaviours commonly seen in nursing home residents. Behaviours are characterized in four clusters: verbally aggressive (e.g., directed at a person or object), verbally nonaggressive (not directed at a specific object or person), physically aggressive (directed), and physically nonaggressive (undirected). The total score is most commonly used to quantify behavioural disturbance. Professional caregivers, who are usually nurses or nursing assistants, assess this scale. The staffs are trained prior to using the instrument. Factor analysis has shown three basic dimensions underlying the 29 CMAI items: physical aggression, physical nonaggression, and verbal agitation. It is a reliable and valid instrument to measure behavioural symptoms in dementia patients [
49,
50].
The NPI-NH is a structured interview with a healthcare professional (CNA/RN). In this interview ten neuropsychiatric symptoms are assessed: delusions, hallucinations, agitation/aggression, dysphoria, anxiety, euphoria, apathy, disinhibition, irritability/lability. Screening questions are asked to determine whether behavioural changes are present. In the case of a positive answer, further questions are asked and the severity and frequency of the behavioural disturbances are determined. The Dutch version of the NPI has high inter-rater agreement and is found to be a valid rating scale for measuring a wide range of behavioural and psychological symptoms of dementia [
51,
52].
Pain
Symptoms of pain will be measured with the pain scale of the Dutch version of the Minimum Data Set of the Resident Assessment Instrument (MDS-RAI) [
34] and with the Pain Assessment Checklist for Seniors (PACSLAC-D) [
37]. The MDS contains two pain items: 'pain frequency' and 'pain intensity'. In the MDS, pain frequency is coded as no pain (0); less than daily pain (1); and daily pain (2) in the last 7 days. Pain intensity is categorized as no pain; mild pain (0); moderate pain (1); and severe pain (2, times when pain is horrible or excruciating) in the last 7 days. The product of intensity and severity in this study has to be ≥2 to be defined as clinically relevant pain.
The validity and precision of pain measurement with MDS items have been established against the Visual Analogue Scale in a study involving 95 US nursing home residents [
34]. The definition of pain in the MDS is ''Pain refers to any type of physical pain or discomfort of the body. Pain may be localized to one area, or be more generalized. It may be acute or chronic, continuous or intermittent (comes and goes), or occur at rest with movement. The pain experience is very subjective; pain is whatever the resident says it is." Coding instructions are ''Code for the highest level of pain present in the last seven days" [
54].
The PACSLAC-D is a brief and manageable version of the Pain Assessment Checklist for Seniors with Limited Ability to Communicate [
35] in Dutch that was developed by Zwakhalen [
37], with a three component solution including 24 items. This version had high levels of internal consistency for the complete scale (Cronbach's alpha range 0.82-0.86) and for all subscales (alpha range 0.72-0.82).
Both the pain scale of the MDS-RAI and the PACSLAC-D will be assessed by the contact nurse of a particular resident at baseline and after 3 and 6 months.
Video registrations
At baseline, 5 minute standardized video recordings of the included patients will be made during both morning care and mealtime. This will be repeated after three months.
Independent, blinded raters will score the video-recordings, by using the PACSLAC-D. These video-recordings will also be used to independently code another pain and discomfort scale, the Dutch version of the Discomfort Scale- dementia of Alzheimer Type (DS-DAT) [
26,
55]. The DS-DAT is an observational scale with 9 items, and measures symptoms regarding vocalizations, breathing, facial expression, and body movement. The reliability and validity of the Dutch version has been established as high [
55‐
57]. In addition, the video-recordings will be analyzed using the Facial Action Coding System (FACS) [
58].
Demographics, baseline and control variables
Sex, age and cultural background of residents will be derived as demographic variables from structured questionnaires filled in by nurses. A subscale of the Residents Assessment Inventory (RAI) will be used to obtain co-morbidity and will be filled in by the nursing home physicians. They will also determine the stage of dementia, measured with the Global Deterioration Scale [
29,
62] and the Cognitive Performance Scale (MDS-CPS).
The MDS-CPS is a seven-category index, ranging from cognitively intact to very severely impaired. The scale has shown excellent agreement with the Mini-Mental State Examination (MMSE) in the identification of cognitive impairment in research [
63].
The Global Deterioration Scale (GDS) [
29] consists of a seven-point scale (1-7) ranging from 'no global impairment' (1) to 'very severe global impairment' (7). Characteristics of the nursing homes (such as available disciplines, amount and nature of followed education by health care professionals, unit sizes) will be carefully registered and taken in to account when performing the analyses.
Medication
Changes in prescriptions of analgesics and psychotropic drugs (coded as defined daily dosage-DDD) will be derived from the daily logs of contact nurses and will also be derived from the analysis of pharmacists' electronic patient records for the total study period of 26 weeks.
Apolipoprotein Epsilon 4
The status of Apolipoprotein Epsilon 4 (Apo-E4) will be determined as a baseline variable and will be taken from two buccal mucosa swabs (Catch-All swabs of the firm BIOzymTC example). The Apo-E4 protein may play a role in the response to (pharmacological) treatments [
64]. It is currently unclear whether the Apo-E4 protein has an effect on non-pharmacological treatments, such as comfort interventions. Determining the Apo-E4 is therefore very important in intervention studies like the present study.
Process variables
The actual use of the STA OP!-protocol will be derived from the daily logs of contact nurses for the total study period of 26 weeks, and from group interviews with contact nurses and nursing home physicians in weeks 26. The researcher/research assistant making site-visits once a week, will check the thoroughness of the log completion by the contact nurse. After the research period the internal formed focus group will take over this task, in order to facilitate implementation.
Patient files
The use and nature of non-pharmacological comfort interventions, as well as the nature and number of physical restraints will be derived from daily logs by the contact nurses and by studying patient files for the total study period of 26 weeks.