1. Background

In Korea, approximately 89.2% of the elderly population is reported to have one or more chronic diseases, and the majority of older adults report experiencing pain [1]. Elderly residents in long-term care hospitals and facilities suffer mainly from diseases which cause pain or require pain management, such as cerebrovascular disease, dementia, and cancer [2]. In addition, most older adults have musculoskeletal disorders, which are the most common cause of pain, and neurological disorders and recent fall experiences, urinary tract infections, pneumonia, and skin wounds are also reported as common causes of pain in older adults [3, 4]. Elderly residents in long-term care hospitals, including dementia patients, suffer from chronic illnesses or physiological degenerative conditions. Pain is one of the most common symptoms they experience during long hospital stays. Actually, in a previous study, more than 80% of patients in a long-term care hospital reported having pain, and the average pain intensity of the participants was rated as moderate or severe [5]. In particular, 66.7% of the non-dementia group and 63.2% of the dementia group reported pain [6], showing that a large proportion of older adults with dementia also report moderate to severe pain, contrary to general expectations.

However, due to the inadequate awareness about pain in older adults among older adults and healthcare workers, pain in older adults is not properly assessed and managed. In addition, although the levels of pain intensity in older adults with dementia are not much different from those of pain intensity in older adults without dementia, pain assessment and prescription of analgesics by medical staff have been reported to be significantly lower in older adults with dementia than in older adults without dementia [7]. Dementia patients experience a decline in linguistic expression ability and communication skills due to cognitive impairment, so it is difficult for medical staff to assess pain in dementia patients because they cannot express their pain adequately. This untreated pain leads to a decline in physical function, sleep disorders, reduced socialization, increased uses of medical services and an increase in medical costs, ultimately deteriorating the quality of life in older adults [4].

On the other hand, it has been found that older adults, including elderly dementia patients, tend to think of pain as a natural phenomenon due to aging, and are reluctant to report pain not to bother medical staff, and have fear about addiction to pain medication. Medical staff regard pain in older adults as a part of the aging process rather than a health problem, and think that dementia patients do not experience pain or cannot report it even when they experience it. Moreover, in some cases, medical staff are not aware that changes in the behavior of dementia patients can be signs of pain [8, 9]. In addition, it has also been reported that medical staff have fear about administering narcotic analgesics to older people, show concerns about drug addiction and drug tolerance in older people, and lack confidence in implementing non-pharmacological pain management [10].

This lack of knowledge of medical staff about pain becomes a major obstacle to pain management, leading to inappropriate pain management [8]. Among healthcare workers, nurses are personnel who provide care for patients around the clock in the nearest position from them. For this reason, they need to be equipped with the competency to recognize pain and provide appropriate pain management. Nurses working in long-term care hospitals, where almost all patients are older adults, should implement pain management with accurate knowledge about pain in older adults, including those with dementia. Nurses without adequate knowledge about pain may not adequately recognize pain in patients, so patients' complaints of pain are ignored and not recorded, making it difficult to perform pain management, and even if pain assessment is performed, a lack of knowledge about pain management may lead to inappropriate pain management [9].

Therefore, for effective management of pain in older adults, nurses should have an adequate understanding of the nature of pain in patients and accurate knowledge of pain assessment and management. A lack of knowledge limits healthcare workers' abilities to understand information related to diseases and perform treatment, leading to missing the appropriate timing for treatment [11]. Nurses should have knowledge about drug addiction and dependence, adverse drug reaction, pharmacological pain management, non-pharmacological pain management, and methods of measuring the effectiveness of pain management as well as knowledge about nonverbal cues of pain or signs and symptoms of pain in older people, including those with dementia [7]. To appropriately manage pain in older people, including those with dementia, it is necessary to accurately assess the knowledge levels of nurses and implement training to improve pain management. To do so, a measurement tool is needed to assess nurses' knowledge about pain assessment and management.

In Korea, there is an instrument developed to assess nurses' attitudes towards pain in patients with severe dementia [12], but there are no instruments developed in Korea or Korean-translated versions of instruments for assessing nurses' knowledge about pain in older adults, including dementia patients. In foreign countries, there are measurement tools developed to assess knowledge and attitudes [9] or knowledge and beliefs [13] regarding pain in dementia patients among nursing staff, including nurses working in nursing homes or to assess knowledge about pain in older adults among nurses working in hospitals and nursing homes [14]. However, these tools involve an ambiguous mixture of knowledge and attitudes or a mixture of knowledge and beliefs, and some aspects of them are not suitable for assessing knowledge alone. In addition, each of them was developed for a single author's own use, and the evaluation of the reliability and validity of the tools was not properly carried out. Therefore, Fetherstonhaugh et al. [7] developed the Pain in Older Adults Knowledge Survey (POAKS) to measure nurses' knowledge about pain in older adults, including dementia patients. This tool is consists of items about the experience, assessment, and management of pain in older adults and the items were developed based on a review of research literature. With respect to the evaluation of the reliability and validity of the tool, 17 experts, including authors of guidelines or research papers regarding pain in older adults and elderly dementia patients, developers of existing pain assessment tools, and members of an international pain society, participated in the evaluation processes and established the reliability and validity of the POAKS through the three-round Delphi process, demonstrating that the POAKS is an appropriate tool for measuring knowledge about pain in older adults.

Therefore, in this study, we aimed to develop the K-POAKS using the POAKS created by Fetherstonhaugh et al. [7] through translation, back-translation, reconciliation, and adaptation to create a tool suitable for Korean culture and to evaluate the reliability and validity of the K-POAKS through a survey among nurses of long-term care hospitals. An assessment tool developed through the instrumentation process described above is expected to provide basic data for the improvement of nurses' knowledge about pain in older adults and implementation of appropriate pain assessment and interventions.

2. Purpose

This study aimed to develop a Korean version of the Pain in Older Adults Knowledge Survey (K-POAKS) through translation, back-translation, reconciliation, and adaptation using the POAKS developed by Fetherstonhaugh et al.[7] to assess knowledge about pain in older adults including dementia patients among nursing staff working in nursing homes, and to evaluate the validity and reliability of the developed tool. The specific objectives of this study are as follows:

1. Study design

This study is a methodological study to develop a Korean version of the Pain in Older Adults Knowledge Survey (K-POAKS), a tool to assess knowledge about pain in older adults among nurses, and to verify the reliability and validity of the K-POAKS.

2. Participants

In this study, nurses working in long-term care hospitals across the country were the target population, but nurses working in long-term care hospitals located in B, P and U cities were selected as the accessible population and the participants were selected by the convenience sampling method. The inclusion criteria were long-term care hospital nurses with 3 months or more of work experience in long-term care hospitals and the experience of taking care of dementia patients with pain. The sample size was determined based on the literature suggesting that the sample size for factor analysis should generally be 4 to 5 times the number of variables and that the sample size required for factor analysis is generally 100 to 200 [15, 16]. Thus, a minimum of 120 samples were required in this study, which was 5 times the number of 24 items of the POAKS. A total of 195 participants were enrolled in the study, and the data of 179 people were finally included in the analysis, excluding 16 people because of missing data.

3. Procedures

4. Data Collection and Ethical Considerations

Data collection was carried out from February 10 to May 30, 2019, and this study was conducted after obtaining approval from the IRB of Ulsan College (UC2018009). To collect data, the researcher visited or called long-term care hospitals located in B, D, or U city and explained the purpose and procedures of the study to the head of the nursing department and requested cooperation for data collection. Then, questionnaires were delivered to the head of the nursing department of each hospital by personal visit or by mail. The questionnaires were distributed after receiving written informed consent from participants, and the participants were requested to complete the questionnaires immediately after receiving them and put completed questionnaires in the document envelopes placed at a designated place. The document envelopes with completed questionnaires were retrieved by the researcher or the head of the nursing department.

5. Statistical Analysis

The data of this study was analyzed using SPSS/WIN 22.0, and the reliability and validity of the K-POAKS were estimated.

First, in order to evaluate the reliability of the K-POAKS, internal consistency reliability was assessed using the Kuder-Richardson Formula 20 (KR-20) coefficient. In addition, the ‘α if item deleted’ value, which indicates what the Cronbach's α value will be if a measurement item is deleted, was computed, and corrected item-total correlation (ITC) was also calculated.

Second, to assess the content validity of the K-POAKS, the content validity index (CVI) was calculated by a group of experts.

Third, to evaluate criterion-related validity, Pearson's correlation coefficient was used to determine the degree of correlation between the K-POAKS and the tool for assessing attitudes toward pain management in dementia patients and between the K-POAKS and the tool for assessing performance of pain management in dementia patients.

Fourth, to evaluate the construct validity of the K-POAKS, the Kaiser-Meyer-Olkin (KMO) test for sampling adequacy and Bartlett test were conducted and exploratory factor analysis was also performed. In addition, in the exploratory factor analysis, principal component analysis was conducted using the varimax rotation method in order to investigate the factor structure of the K-POAKS.

1. General Characteristics of Participants

The mean age of the participants was 41.6 years, and people aged 50 or over were 54 people (30.2%), accounting for the largest proportion. For gender, 169 people (94.4%) were females. For marital status, 121 people (67.6%) were married and 58 people (32.4%) were unmarried. Regarding educational level, the participants consisted of 96 junior college graduates (53.6%), 77 college graduates (43.0%), and 6 people with a master's or higher degree (3.4%). As for position, 147 people (43.0%) were staff nurses and 32 people (17.9%) were charge nurses or in a higher position. The mean period of working in hospitals was 12.35 years, and 66 people (36.9%) worked as nurses for 10~19 years, 45 people (25.1%) for 4 years or less, 38 people (21.2%) for 20 years or more, and 30 people (16.8%) for 5~9 years. The period of working in long-term care hospitals was 3.21 years on average, 1~5 years for 117 people (65.4%), over 5 years for 36 people (20.1%), and less than one year for 26 people (14.5%) (Table 1).

Table 1 General Characteristics of Nurses in Long Term Care Hospitals (N=179) Click for larger image

2. Internal Consistency Reliability

The K-POAKS consists of a total of 24 items, and respondents are required to answer ‘Yes’, ‘No’, and ‘Do not know.’ The responses are scored by assigning 1 point for a correct answer and 0 points for an incorrect answer or ‘Do not know’, and higher scores indicate higher levels of knowledge. The KR-20 coefficient, a measure of internal consistency reliability, was .84 for the original tool. In this study, the KR-20 for all items was .75 (Table 2). Regarding the KR-20 for each factor, the KR 20 was .69 for factor 1, .77 for factor 2, .68 for factor 3, and .68 for factor 4.

Table 2 Content Validity Index, Internal Reliability, Item-total Correlation and Level of Korean Version of The Pain in Older Adults Knowledge Survey (K-POAKS) (N=179) Click for larger image

3. Content Validity

In the evaluation of content validity by experts, the CVI of each of the items of the K-POAKS was 0.84 or higher, indicating that the K-POAKS has an appropriate level of content validity as a tool to assess nurses' knowledge of pain in older adults including dementia patients (Table 2).

4. Item Analysis

In this study, the item-to-total correlation (ICT) was used to determine the correlation between each item and total items. Item 4 (r=.01, p=.847) and item 21 (r=−.09, p=.226) were found to have no significant correlation, but the ICTs of the remaining 22 items ranged from .15 to .65, showing that each of the items has a significant correlation. There were no items with an excessively high contribution rate of .80 or higher.

The mean knowledge score of 22 items, except items 4 and 21, ranged from 0.51 to 0.96. If the mean knowledge score is around 0.5, it indicates that approximately 50% of the respondents answered the item correctly. The mean knowledge score was about 0.5 in 2 items, about 0.6 in 2 items, about 0.7 in 2 items, about 0.8 in 9 items, and about 0.9 in 7 items, indicating the items were slightly easy in terms of the difficulty level. However, for items 4 and 21, which were shown to have a low ICT, the mean knowledge scores were 0.26 and 0.08, respectively, showing that they were very difficult questions (Table 2).

5. Criterion Validity (Concurrent Validity)

The analysis of the relationship between the K-POAKS and the tool to measure nurses' attitudes towards pain management in dementia patients and between the K-POAKS and the tool to measure performance of pain management in dementia patients revealed that the K-POAKS has a significant positive correlation with both the attitude measurement tool (r=.28, p<.001) and the performance measurement tool (r=.21, p=.004) (Table 3).

Table 3 Pearson Correlation Coefficient between Attitude, Performance scale and Korean Version of The Pain in Older Adults Knowledge Survey (K-POAKS) Scale (N=179) Click for larger image

6. Construct Validity

The univariate analysis for each item was performed to investigate if each item meets the assumptions or criteria for factor analysis. The results showed that there were no items that did not have a normal distribution. In addition, the ranges of the skewness and kurtosis of each item did not exceed ±2, indicating that the assumption of normality was satisfied. To confirm whether the items were appropriate for factor analysis, the Kaiser-Meyer-Olkin (KMO) test for each of the 24 items was performed, and the KMO value was .77. In addition, in the Bartlett test, the correlation matrix for the 24 items was χ²=1,120.35 (p<.001), showing statistically significant differences, so the items were shown to be suitable for factor analysis (Table 4). The communality indicates the percentage of variance explained by the extracted factors. If the communality value of an item is below 0.4, it is considered desirable to exclude the item from factor analysis [20]. In the present study, since there were no items with a communality value of less than 0.4, all the items were included in the factor analysis. In the exploratory factor analysis, principal component analysis was conducted using the varimax rotation method in order to assess construct validity, and as a result, seven factors with eigenvalues of 1.0 or higher were extracted. Among the seven extracted factors, except for factors 5, 6, and 7, each of which consists of one item, the other factors were named as follows: factor 1 was named pain assessment and treatment, factor 2 methods of pain expression in older adults, factor 3 pharmacological pain therapy and misconceptions about pain in older adults, factor 4 objective pain assessment in older adults. Regarding the factor loading, which indicates the degree of correlation between each variable and the factor, except for three items of items 1, 4, and 21, factor loadings for the 21 items were all significant levels of .40 or higher. A total of 7 factors were found to account for 58.5% of nurses' knowledge of pain in dementia patients (Table 4).

Table 4 Korean Version of The Pain in Older Adults Knowledge Survey (K-POAKS) Factor Loadings, Eigenvalue and Explained of Variance (N=179) Click for larger image