Implementation of observational pain management protocol to improve pain management for long-term institutionalized older care residents with dementia: study protocol for a cluster-randomized controlled trial

The long-term care homes, funded by the government, must be under the Social and Welfare Department of Hong Kong (HK) and thus meet a set of standard criteria as set by the HK Government. The staffing pattern, staff-resident ratio, bed space, the other service criteria of the facilities, and the nature of the residents are therefore similar in that they have to meet government standards. Additionally, the total number of residents must be 100 or above so as to maximize the chance of recruiting a sufficient number of participants in each care home (that is 15 to 16 - please refer to sample size for details).

Private nursing homes will be excluded due to the variations in their staffing pattern, the staff-resident ratio, bed space, the other service criteria of the facilities, and the nature of the residents.

The target population of this study is long-term care residents with dementia, communication impairment, and pain problems.

Participants must be ≥65 years old, of either gender. They must have been living in the long-term care home for >6 months to avoid anxiety and distress due to the unfamiliar new environment. Their general health must be stable (with no acute physiological or psychiatric illnesses, and no history of fractures or emergency hospital admissions within the past 3 months). Participants must be officially diagnosed with some form of dementia according to the DSM-IV criteria for dementia. Their cognitive state will be assessed by the Cantonese-Mini Mental State Examination (C-MMSE) [20]. The participants must have at least one pain-related diagnosis such as osteoarthritis, low back pain, or rheumatoid arthritis. Each participant’s communication ability will be assessed by the communication score based on two items in the Hong Kong version of the interRAI-Home Care (HC) Assessment (formerly called the Minimum Data Set-HC): (1) making one’s self understood; and (2) ability to understand others. These two items are rated on a 5-point scale: 0 = understood, 1 = usually understood, 2 = often understood, 3 = sometimes understood, 4 = rarely or never understood. A total communication score is summed from these two items and ranges from 0 to 8, with 0 meaning no communication problems and 8 meaning very severe communication impairment. Participants must have a communication score ≥5 to be eligible for recruitment into this study [21].

Exclusion criteria includes residents with: (1) a current medical condition and the need to be frequently admitted to hospital during the study period (for example, acute exacerbation of chronic obstructive pulmonary disease or renal failure, and so on); (2) co-morbid psychiatric disorders (for example, psychosis); or (3) a recent experience of distressing social circumstances (for example, the death of a close relative or friend within the previous 6 months) that may cause altered behavior patterns; and (4) terminal illness (for example, advanced cancer) and deemed not to have more than 6 months to live.

Although many kinds of chronic diseases such as diabetes mellitus and ischemic diseases may affect the pain sensation of older people, it is not uncommon for them to have multiple health problems. Residents who match all the inclusion criteria but have other kinds of chronic illnesses will still be recruited, providing the illnesses are not acute. Each participant’s medical history will be recorded in detail.

We will randomly select eight long-term care homes that agree to join this study. As no similar cluster-randomized trial has been reported in the literature, we can only estimate the intracluster coefficient with a range between 0 and 0.1 [22]. The intensity of pain can vary according to the different natures and sources of the pain. As this study is not limited to participants suffering from a particular type of pain, we anticipate that variations in the nature of participants’ pain will be considerable. Due to its heterogeneous nature, we assumed an intracluster correlation of 0.02. The pilot study was able to detect a mean difference of 13.41 points on the MSQ score (95% confidence interval 8.923 to 17.897) and -0.3 points on the C-PAINAD (95% confidence interval -4.459 to 0.169) between the pre- and post-tests [18]. After averaging the effect sizes calculated based on these two mean differences, this was equivalent to an effect size of 0.65. In order to secure an appropriate effect size, a more conservative approach of using an effect size of 0.4 was used to calculate the sample size. With a significance level (α) of 0.05 and power (1-β) 0.8 for a two-sided test, the sample size calculation gives about 38 participants for each group in order to evaluate a mean difference in MSQ and C-PAINAD scores between the experimental and control groups. Accounting for the intracluster correlation, a variance inflation factor of 1.31 will be multiplied to this calculated sample size. Further accounting for a 15% drop-out rate over the period based on two similar studies [13, 18], a total number of 61 participants per group is required, resulting in a total proposed sample size of 122. This means we will recruit 15 to 16 participants in each long-term care home.

Baseline assessment will include the weekly MQS III score and the weekly record of non-pharmacological pain treatments. Participants’ daily use of pain medications will be quantified into the MQS III scores. Other non-pharmacological pain-relieving interventions will also be recorded daily. Both of these records will be collected weekly for the first 4 weeks of the intervention cycle by the designated research assistant (RA). This collection will continue for the 12 weeks of the Protocol implementation in the experimental care homes or during the usual pain management period in the control care homes. In addition, participants’ C-PAINAD scores will be collected by the independent pain observer once a week throughout the 16-week study period. Please refer to the section below for a detailed explanation of the role of the independent pain observer.

Four 1-hour workshops will be conducted by JL during the first 4 weeks of the intervention cycle. These will aim to ensure that both professional and non-professional care home staff understand and follow the Protocol correctly. Workshop I will introduce pain assessment and management for older people with dementia. Workshops II and III will explain the use of the Protocol for their pain assessment and treatment. The staff will practise pain assessment by C-PAINAD through various pre-developed clinical vignettes. They will also learn to discriminate between various steps of pain treatments based on the pain score. Group discussions will be held after the practice. The staff will be asked to explain their choice of indicators and give reasons for their choice. They will also be asked to explain their subsequent actions in terms of the selection of different stages of pain treatments suggested by the Protocol. Through practice and discussion sessions, the staff should be able to follow the Protocol. Workshop IV is mainly about the implementation of the Protocol. Finally, a test will be conducted to ensure that the staff understand the Protocol. Given that the job nature and educational backgrounds of professional and non-professional caregivers are different, they will attend the workshops separately. The workshops’ contents will be tailored to suit these two groups of staff.

Following the training workshops and after ensuring that all care home staff understand and can follow the Protocol (Table 1), it will be formally incorporated into the care homes’ daily routine.

The staff in the experimental care homes will be aware of the pain treatment received by the participants. In view of the fact that possible expectancy bias may develop among staff in the experimental group as the study progress, the observational pain scale C-PAINAD will also be completed by an independent pain observer, who will be blinded to the study’s hypotheses, the allocation of participants to the experimental/control conditions, and their total pain treatment consumption. As mentioned, the pain assessment will be conducted weekly during the 16-week study period (baseline assessment for 4 weeks and Protocol implementation for 12 weeks) in both the control and experimental homes. Participants will perform the standardized exercise programme while the pain observer is assessing their pain. In previous studies, the standardized exercise programme successfully triggered pain-related behavioral responses [32, 33]. The exercise programme will include a set of passive range of motion (ROM) exercises involving the movement of all major joints. Each ROM will be repeated 10 times or moved to the point of resistance and held for 10 seconds. The participants will then be asked to perform a set of active motions that include changing position from lying to sitting, from sitting to standing, and then from standing to walking. The design of this standardized exercise programme was based on a literature review and followed consultation with a physiotherapist and a nurse with long-term care experience. This exercise programme is similar to the daily stretching exercises performed by physiotherapists in care homes. It helps older people retain joint flexibility and avoid joint stiffness. It was specially designed and has been used to evaluate the psychometric properties of four observational pain scales, including PAINAD [34]. It is believed that if participants suffer from pain under routine nursing care activities, this exercise programme will generate a level of pain similar to that experienced by the participants. This exercise will be performed by another well-trained RA who will be blinded to the study hypothesis and group allocation.

As in the experimental long-term care homes, four 1-hour workshops will be conducted by JL, addressing pain assessment and management for older people, during the first 4 weeks of the intervention cycle. The control care homes will continue their usual pain management strategies for the participants for the remaining 12 weeks of the intervention cycle. Similarly, participants’ weekly C-PAINAD scores will be collected by the independent pain observer throughout the 16-week study period. Their weekly MQS III score and weekly record of non-pharmacological pain treatments will be collected by another RA throughout the intervention cycle.

The MQS III will quantify the usage of pain medications [19, 35]. The MQS I was used as an outcome indicator to demonstrate that regular use of an observational pain tool could significantly increase the total dosage of pain medications [13]. The MQS quantifies pain medications according to their daily administered dosage, pharmacological classification, and detriment weight (that is, the potential of each medication to produce side-effects) [19]. The concurrent validity of the MQS I was established by calculating the correlation coefficient (r = 0.755, P <0.01; two-tailed) between MQS I scores and the mean clinical judgment of pain management professionals [19]. The MQS III was updated in 2003 to modernize medication classifications and their detriment weights [36]. The MQS III has been validated and applied to the study of pain regimens in various pain conditions [37‐39].

The common non-pharmacological interventions used for relieving pain among older people include, but are not limited to, heat and cold therapy, massage, transcutaneous electrical nerve stimulation (TENS), and acupuncture. Only interventions initiated by healthcare professionals will be recorded. The type and the total time per week of the treatments will be recorded during the study period. The non-pharmacological pain treatment will be recorded weekly for 4 weeks by the same RA, who is also responsible for collecting MQS III scores, according to participants’ nursing, physiotherapy or occupational therapy records prior to Protocol implementation (baseline period) and for the 12 weeks of Protocol implementation.

The C-PAINAD consists of five pain behaviors: breathing, negative vocalization, facial expressions, body language, and consolability. Each behavior is rated from 0 to 2 according to the severity of behavior exhibited, thus giving a total score of 0 to 10. Construct validity was supported by the C-PAINAD’s ability to differentiate pain from other pain-free conditions. It can detect pain among people with dementia. Exploratory factor analysis was used to evaluate the construct validity in which one major factor (that is, pain) was extracted and the percentage of variance was 51.2% [40]. The correlation coefficient was 0.68 when comparing the C-PAINAD with the Pain Visual Analog Scale. Generally, the findings were similar to those of the initial study validating the original PAINAD [28]. The results showed that the C-PAINAD was a simple, reliable, and effective pain assessment tool for cognitively impaired older people. A well-trained independent pain observer (that is, another research assistant who will not be involved in collecting data related to pain treatments) will use the C-PAINAD to observe participants’ pain scores weekly during the 4-week baseline assessment period. This will then continue weekly for 12 weeks during the Protocol implementation. Nurses in the experimental condition will also use the C-PAINAD to observe the pain-related behavior of participants daily for the 12 weeks of Protocol implementation. Pain scores collected by nurses will be compared to those collected by independent pain observers.