Psychometric Properties of the ‘Patients’ Perspective of the Quality of Palliative Care Scale’

INTRODUCTION

Caregiving is one of the fundamental components of health system services^[1] and providing good quality care, which is constantly evolving, is a priority in every health system.^[2] Palliative care is one of the care types in the clinical setting.^[3] It is a basic human right for all patients, especially those requiring constant care due to chronic, progressive and excruciating pains caused by chronic malignant and non-malignant diseases.^[4] Palliative care is comprehensive and holistic care provided to patients by an interdisciplinary team of physicians, health-care providers, nurses, psychologists and social workers during hospitalisation and even after discharge.^[5] Palliative care is aimed to provide physical, psychological, social and spiritual support to patients with chronic diseases and those in the end stages of life as well as their families through the prevention of suffering.^[6] Such care is always required by patients with chronic progressive diseases, limitation in daily activities, uncertain course of the disease, instability in symptoms and disruptions in the process of life.^[7]

Reports reveal that about 40 million chronic patients in the world with cardiovascular disease (38.5%), chronic lung disease (10.3%), AIDS (5.7%), diabetes (4.6%) and cancer (37%) annually receive palliative care.^[8] Providing palliative care has positive consequences such as an increase in the patient’s energy, pain relief, dealing with the disease taboo, changing the lifestyle^[9] and promoting the patient knowledge and awareness.^[10] However, it can be impaired by challenging factors, including inefficient performance, poor or non-quality care or care that does not meet the standards,^[11] progression of the destructive power of disease,^[12] progressive dysfunction of daily function,^[13] heterogeneous coverage and improper level of services^[14] and inadequate access to care services.^[15]

Meier (2011) showed that chronic patients hospitalised in different wards do not receive optimal-quality palliative care services. This is manifested by untreated symptoms and unmet social, psychological and physical needs, gradually resulting in increased caregiver burden and decreased patient satisfaction.^[16] The results of the qualitative research conducted by Ansari et al. also revealed that the quality of palliative care has some challenges in Iran and needs to be constantly evaluated.^[17]

According to the National Institute of Health Care Quality, evaluation and monitoring strategies should be considered to achieve high-quality care in the community.^[18] Evaluating the quality level of care programs regarding the opinions of stakeholders and care receivers is an efficient strategy.^[19] Furthermore, research shows that developing palliative care always depends on understanding the patients’ needs and their perspectives on the physical, psychological, social, spiritual and cultural quality of care and continuous evaluation of the process.^[20] In other words, an adequate insight and clear picture of palliative care quality will be achieved through getting the patients’ opinions and their real experiences.^[21]

Various palliative care quality assessment instruments have been developed throughout the world to gather the patients’ perspectives on the quality level of palliative care. For instance, Yun et al. developed a questionnaire to assess the quality of palliative care in cancer patients in Korea,^[22] and Guirimand et al. devised and validated a questionnaire on the patients’ perspectives of the quality of palliative care in France.^[20] Moreover, Buzgova et al. designed an instrument to assess the patients’ needs in palliative care and levels of care.^[23] By the same token, Lyons et al. validated the Functional Assessment of Chronic Illness Therapy-Palliative care scale to appropriately assess the quality of existing care.^[24]

However, there is no instrument in Iran to assess the quality of palliative care provided to patients. Thus, considering the progressive trend of providing palliative care, local and specialised instruments is required for assessing its quality. Such instruments must follow the cultural factors, facilities and conditions of care provider organisations, treatment and care team skills in this field. Using such instruments provides the opportunity to accurately evaluate various dimensions of the quality of palliative care in patients while addressing its shortcomings and challenges. Hence, the present study was developed and conducted to determine the psychometric properties of the patients’ perspectives of the quality of the palliative care scale.

MATERIALS AND METHODS

RESULTS

From 500 patients participating in the study, the majority of the participants in the study were men (263, 52.5%), married (405, 81%) and illiterate (286, 57.2%) with a mean age of 55.38 ± 15.51. Totally, among the chronically hospitalised patients, 376 (75.2%) suffered from non-malignant and 124 (24.8%) were affected by malignant diseases.

In the initial reliability assessment, the results of the adjusted Cronbach’s alpha revealed that items 7, 11. And 26 should be deleted. By omitting these three items, there was an increase in the total Cronbach’s alpha value of the scale and subscales [Table 1].

In qualitative face validity assessment, ambiguities posed by the target group in understanding 12 items (i.e., 13, 5, 14, 15, 16, 17, 9, 20, 21, 31, 34 and 35) were resolved by the explanations of the researcher. Ultimately, the qualitative face validity was confirmed. Calculating the impact factor score, all items obtained an impact factor of 1.5 and so remained in the scale.

In evaluating the qualitative content validity, five items (1, 9, 13, 22 and 28) were revised. Moreover, the Likert scale was altered to ‘Does not apply to me’ (0), ‘Never’ (1), ‘Hardly ever’ (2), ‘Sometimes’ (3) and ‘Always’ (4) based on the nature of the questions. The values of the items were confirmed using the Lawshe table (>0.62). Scale-CVI/Ave value was 0.94 and the indices of all items were reported within 0.67–1.

For the construct validity, participants who were literate completed the questionnaire independently and without the need for assistance. Participants without literacy answered the questions with the help of the researcher. Out of 500 samples collected through simple random sampling, 200 samples were allotted to EFA and 300 samples to CFA. In the EFA group, exploratory factor analysis was conducted considering the KMO index and Bartlett test. The KMO index value of 0.863 (i.e., > 0.6) indicated that the number of samples was adequate for factor analysis. The statistical significance of the Bartlett test (i.e., 0.00) revealed that factor analysis was appropriate to identify the structure of the factor model.

Then, accepting a factor load of ≤0.3 and Eigenvalue>1, the EFA was performed on eight subscales with 32 items [Table 2].

Based on the EFA results in [Table 2], after rotating the factor loadings, 32 items were placed in seven subscales. Besides, the subscale of ‘possibility to refuse (care or volunteer)’ was removed and its items were moved to other subscales. Quali-Palli-P explained 63.94 of the total variance. The highest variance was allocated to the factor of ‘availability of the caregivers’ and the lowest variance was related to the factor of ‘respect for the patients’ [Table 3 and Figure 1]. Considering the structure of the Quali-Palli-P model, CFA results with AMOS 24 software indicated that items 10, 32 and 24 were also removed from the scale. The fit indices of the obtained model properly confirmed the omission of three factors. Therefore, the number of items was reduced to 29. Fitness indices revealed there was a good fitness between the model in the data of this study and the obtained factor structure. Hence, Quali-Palli-P with 29 items and seven subscales were finally approved [Figure 2 and Table 4].

Close

Figure 1:

Scree plot in exploratory factor analysis.

Export to PPT

Close

Figure 2:

The measurement model for the interpretation of data obtained from integration based on standard factor loadings F1: Availability of caregivers, F2: Serenity, F3: Providing cure and care information and the patient’s involvement in decisions, F4: Pain management, F5: Caregiver’s listening skills, F6: Psychosocial and spiritual dimensions, F7= Respect for the patients.

Export to PPT

Among the 35 items in the original scale, six items were omitted and Quali-Palli-P factor structure was confirmed with 29 items and seven subscales including ‘availability of the caregiver’ (six items), ‘serenity’ (three items), ‘providing cure and care information and the patient’s involvement in decisions’ (four items), ‘pain management’ (three items), ‘caregiver’s listening skills’ (six items), ‘psychosocial and spiritual dimensions’ (three items) and ‘respect for the patients’ (four items).

The results of convergent validity of Quali-Palli-P showed that the AVE of all factors (0.521-0.610) was >0.5 with the CR of each factor being higher than its AVE [Table 5]. Furthermore, the reliability results revealed that Cronbach’s alpha and McDonald’s omega coefficient of the scale were confirmed. The internal consistency of the total scale was 0.88 and it ranged from 0.71 to 0.93 for the subscales. The ICC of the scale was approved through test-retest (95% confidence interval, r=0.89, significance level of 0.00). It was 0.89 for all items and 0.85-0.88 for the subscales [Table 5].

DISCUSSION

Palliative care is performed not only to improve the quality of life but also to prevent or relieve the suffering of patients and their families.^[34] According to the results of the present study, Quali-Palli-P was introduced as a valid and reliable scale in the Iranian Palliative care context.

In the psychometric process, first, the scale was translated based on the approach posed by Wild et al.^[25] Translating the instrument is one of the most common methods for cultural adaptation or localisation. Therefore, paying attention to the proper translation process and cultural adaptation of meanings leads to validity the reliability of the instrument. Researchers believe that there is no specific guideline for translating the instruments and the stepwise process of translating and cultural adapting based on a specific pattern brings about high-quality results.^[35]

Some researchers believe that meeting four standards, including content validity, construct validity and two types of reliabilities are necessary for validating the instruments and making them applicable in the research context.^[36] In the present study, the stages of initial reliability, face, content, convergent and construct validities were precisely performed. Accurately performing the initial reliability and studying its results facilitates a psychometric path.^[37] Moreover, face validity was confirmed by obtaining the comments of the target groups, making minor changes on some items and calculating the impact score. Polit have stated that face validity assessment supports the ability of the target group to comprehend the items of an instrument.^[38]

The qualitative content validity of the scale was also confirmed by revising several items in a panel of experts using a Likert scale. Moreover, the CVR and CVI of all items were verified and the items were preserved. Experts believe that in content validity assessment. The more knowledgeable the selected experts in the field, the more accurate the content validation process will be.^[38]

The initial results in EFA indicated the desirability of KMO tests (0.86). Being significant, Bartlett’s test of sphericity revealed the correlation between the studied variables and the correct application of factor analysis. The sample size adequacy in EFA requires values >0.6 in the KMO test.^[39] The items also explained 63.84% of the total variance. In this regard, Zikmund et al. reported that it is appropriate to explain about 60% of the total variance in the domains.^[40] Factor analysis is one of the reliable techniques for validating the instrument; hence, the instrument can be modified by construct validity.^[41]

The EFA results led to the removal of the subscale of ‘possibility to refuse (care or volunteer)’ and moving its items to other subscales. In this regard, Guirimand et al. claimed that the ‘right to refuse some treatments’ was the only dimension that was not correlated with the satisfaction of the quality of care and did not affect it directly.^[20] In the present study, the highest variance was attributed ‘availability of caregiver’ and the lowest belonged to ‘respect for the patients.’ The ‘availability of caregivers’ remained in the scale with high variance, suggesting that patients consider this factor effective and important in the quality of the provided care, whereas the role of ‘respect for the patients’ was poor in care quality assessment. Moreover, Van Soest-Poortvliet et al., Teno et al., Aspinal et al. and Stewart et al. emphasised the higher possibility of removing or moving the items under subscales with a poor contribution in explaining the variance.^[42-45]

Subsequently, a structural model consisting of 32 items in seven subscales was extracted through the EFA and CFA. These subscales included ‘availability of caregivers,’ ‘serenity,’ ‘providing cure and care information and the patient’s involvement in decisions,’ ‘pain management,’ ‘caregiver’s listening skills,’ ‘psychosocial and spiritual dimensions’ and ‘respect for the patients.’ The result has shown that it is essential to calculate EFA before CFA; these two stages of factor analysis are complementary. Following up EFA through CFA is one of the common approaches to validate construct equations.^[30] However, in a similar study conducted by Sandsdalen et al. on patient’s preferences in palliative care, only EFA was performed.^[46]

By removing three factors, the CFA and GFI outputs led to a scale with 29 items in seven subscales. Similar to the results of Dy et al., Guirimand et al. and Vedel et al., ‘availability of the caregivers’ and ‘serenity’ were two important subscales of Quali-Palli-P and the essential components of palliative care quality.^[47-49] Another subscale was ‘providing cure and care information and the patient’s involvement in decisions,’ which was emphasised in the studies conducted by Ansari et al., Khoshnazar et al. and Jabbari et al.^[17,50,51] Moreover, Teno et al. indicated that patient’s participation in care-related decision-making caused a higher quality of care.^[43] Pain management, as well as proper prescribing and applying treatment and care techniques, is other important dimensions of palliative care.^[52] Chronic pain of patients is one of the most important issues requiring palliative care in all physical, psychological, emotional and spiritual dimensions that can affect the quality of palliative care and the life of patients.^[53]

The subscales of ‘caregiver’s listening skills’ along with ‘respect for the patient’ are the cornerstones of standard palliative care through communicating and paying attention to the patient in the process of care. Shahvaroughi Farahani et al. highlighted listening skills of the treatment team and paying attention to the patient and respect for the patients’ desires in palliative care.^[54] The subscale of ‘psychosocial and spiritual dimensions’ should be noted since researchers believe that the nature and long process of chronic diseases require a palliative care system in psychological, emotional, spiritual and social dimensions and can help patients to live an active life and improve their quality of life.^[55] By the same token, Rego et al. expressed the importance of the psychosocial and spiritual dimensions in relieving patients and more effective therapeutic performance.^[56] In the present study, the convergent validity of the scale showed the correlation of the items. The convergence of the instruments is not confirmed when the hidden factors are not well explained by the extracted items and the items are not sufficiently correlated with each other.^[57]

Cronbach’s alpha, McDonald’s omega, AIC, ICC and CR values confirmed the reliability of the scale. The reliability of the instrument is related to its homogeneity and when the instrument passes through the design and psychometric steps correctly and stepwise, good reliability is normally reported.^[58] According to Plichta et al., when the factors of a model are evaluated in a research population, the model enjoys a high internal consistency.^[59]

One of the advantages of the present study was developing the Persian version of the palliative care quality based on the perspective of the patients, which was obtained consistent with the research community, facilities and the clinical setting of the country. Another positive point was using the comments of both the stakeholders and the care receivers on the clinical accreditation indices and the patient’s right charter. On the other hand, one of the limitations of present study is building the different structural equation models using the same data, which can lead to different interpretations.

CONCLUSION

Quali-Palli-P as a specific scale with 29 items and seven subscales can accurately indicate and predict the quality of palliative care from the perspective of chronically hospitalised patients. This scale can be used in assessing the therapeutic issues and quality of palliative care provided to patients in the health system. Furthermore, the patients’ perspective of the quality of palliative care scale, made for the ^1st time in Iran, can be used by all researchers and instructors of medical universities and clinical research centres.