Exploring the Determinants of Patient Satisfaction in Emergency Healthcare: Evidence from Irish EDs

doi:10.21203/rs.3.rs-5010398/v1

Background

Patient satisfaction is a crucial indicator of healthcare quality, particularly in emergency departments (EDs), and it influences both clinical outcomes and institutional reputation. In Ireland, despite the significant challenges faced by these settings, including long waiting times and resource constraints, patient satisfaction in EDs has not been thoroughly explored. This article aimed to determine the key indicators influencing patient satisfaction in an Irish Emergency Department (ED).

Methods

This study was conducted in the ED of an adult teaching hospital in Ireland, utilising a patient satisfaction survey distributed to individuals aged 16 and above who visited the ED. The patient satisfaction determinants were evaluated by estimating multivariate models using PLS analysis to test hypotheses and examine the associations between patient satisfaction dimensions and overall patients’ satisfaction as a dependent outcome measure.

Results

The analysis revealed that patient satisfaction is significantly and positively influenced by the dimensions of information, responsiveness, and assurance. These findings are crucial for understanding and improving healthcare quality. Interestingly, reliability was found to have a negative impact on patient satisfaction. However, it was also found to play a mediating role in the relationship between information and patient satisfaction, highlighting the complex dynamics of patient satisfaction. The dimensions of tangibility and empathy were not significantly associated with overall satisfaction. The expected mediation effect of responsiveness on reliability and satisfaction was not supported.

Conclusion

The findings of this study offer valuable insights for healthcare providers and policymakers. By emphasizing the critical importance of improving the quality of information provided to patients, addressing issues related to reliability, and enhancing responsiveness, this study provides practical guidance for enhancing patient satisfaction in emergency care settings. These factors are crucial in meeting patient expectations and boosting overall satisfaction. Continuous monitoring and targeted interventions are recommended to ensure that patient satisfaction levels are maintained and improved in the dynamic environment of emergency care. Implementing these strategies can lead to better patient experiences, improved clinical outcomes, and a stronger institutional reputation.

Management

Patient satisfaction

Emergency Department

quality of healthcare

partial least squares structural equation modelling (PLS-SEM)

The major shift in the expectations of taxpayers from hospitals has attracted the attention of healthcare decision-makers. Patient satisfaction has also taken on great importance as a measure of quality healthcare delivery. Over the past decade, gauging patient satisfaction has been a growing priority, with great concern for healthcare providers ensuring that patient satisfaction is met (Ferreira et al., 2023). It has also become one of the critical dimensions of quality and part of financial incentives and public reporting requirements (Omaghomi et al., 2024). Excellence in healthcare settings is measured by health outcomes as well as patient satisfaction (Novitasari, 2022).

In today’s competitive healthcare market, managers cannot afford to neglect patient satisfaction scores because more satisfied patients are more likely to recommend the hospital to family and friends (Atsavapranee et al., 2024). Therefore, monitoring patients' perceptions is a fundamental strategy for planning improvement (Han et al., 2020). Hence, healthcare organisations and systems today are under mountain pressure to improve the quality of care provided while minimising costs (Bhat et al., 2020). Furthermore, the healthcare system is considered one of the most complex because of its high variability and uncertainty (Alowad et al., 2020). In response to ubiquitous healthcare complexities, healthcare leaders seek effective tools and methods to provide practical solutions (Okolo et al., 2024) that positively impact patient satisfaction.

Poor patient satisfaction profoundly affects healthcare systems' clinical and financial outcomes. It undermines patient compliance, leading to poorer clinical outcomes, lower care quality, higher readmission rates, and increased medicolegal risk (Sonis et al., 2019) and/or death within 6 months post-discharge (Anderson et al., 2020). For instance, studies show that in hospitals with low patient satisfaction scores, the readmission rate jumped from 5.6% at 30 days to 23.3% at 180 days (Hughes and Witham, 2018).

Over the past decade, a growing global interest in patient satisfaction has been growing. In the United States (US), significant mandatory changes were made to the payment program for all US hospitals. As of 2023, hospital reimbursements are linked to patient experience and the quality of healthcare services and the payment program model is based on four measures: patient experience, the process of service delivery, mortality rates, and readmission rates (Poorani et al., 2023).

As in other service industries, healthcare must be consumer-focused rather than provider-focused (Park et al., 2022). For instance, researchers have suggested that patient- and family-centred care is a mutually beneficial partnership with healthcare providers, patients, and families. As a result, patient input is essential for determining whether patients’ needs are met. From this perspective, it is necessary to view care through patients' eyes to understand their expectations and satisfaction (Alibrandi et al., 2023). However, this has not always been the case, and improvement based on patient feedback rarely materialises (Sheard et al., 2019). Hence, addressing patient satisfaction factors is vital to an organisation's performance. Unfortunately, identifying these factors alone is not sufficient. Once patients' satisfaction is determined, the next critical step is to employ those results to improve the services provided.

Since 2009, patient satisfaction has significantly impacted healthcare hospital funding in the United Kingdom (UK). They must review patients' experiences as part of their quality and performance reporting. Up to 10% of the payments to the National Health Service (NHS) Trusts will eventually depend on reported adequate patient satisfaction levels. The proportion of service funding is contingent on achieving improvements in patient-reported experiences (Department of Health, UK, 2018).

In Ireland, in 2019, the Health Service Executive (HSE) was established on January 1, 2005, by the Health Act of 2004 and officially began operation on January 1, 2005. It took over health services from 11 regional health boards (Saul, 2024). Despite all the efforts made to improve the healthcare system in Ireland, many challenges remain and require extra effort related to long waiting hours, shortages of beds and budgets. Furthermore, the National Care Experience Program was established to improve health and social care service quality by asking people about their care experiences. The program is a partnership between the Health Information and Quality Authority (HIQA), the Health Service Executive (HSE), the Department of Health, and patient representatives (Beecher et al., 2021).

Accordingly, the Irish National Inpatient Experience Survey provides a clear picture of the quality of acute healthcare services for inpatients in Ireland, as reported by patients. However, patient feedback in the emergency department is not utilised to drive improvements, as no survey was developed to explore patients' satisfaction levels in the emergency department in Ireland. Accordingly, this study aimed to examine the determinants of patient satisfaction in emergency departments in acute hospital contexts in Ireland.

Study Setting

The study was conducted in Ireland's Emergency Department (ED) of Adult Teaching Hospital. The selected hospital has one of Ireland's largest and busiest ED. Getting the executive team's approval was crucial to start the project successfully. Accordingly, the initiating phase, which lasted for three weeks, began with a meeting with the hospital's Chief Executive Officer and Chief Operation Officer, followed by several meetings with ED stakeholders, including consultants and head nurses. The project's scope and expected outcome were agreed upon during these meetings. The project challenges were also discussed to run the project and hinder these challenges successfully.

The ED serves approximately 50,000 patients a year and continues to grow in patient volume. Between 2012 and 2023, there was a 22% increase in the annual volume of patients attending the ED without increasing resources and capacity. In 2012, 43576 patients visited the ED, while more than 52000 patients did so in 2023. The satisfaction questionnaire survey was mailed to patients over 16 years visiting the ED, whether they were admitted or not. This survey has 32 questions, with 3 asking about the patient characteristics. The rest of the questions asked the patients to rate their satisfaction with the care and service. The last question asked the patients to rate their overall satisfaction with care.

Survey Instrument

A patient satisfaction questionnaire was developed based on a conducted exploratory study (Swallmeh et al., 2018) and a literature review. The patient satisfaction questionnaire contains 24 questions. The questions asked the patients to rate their satisfaction with the care and service. Patients’ satisfaction statements use a five-point Likert scale ranging from very poor to excellent. Finally, one question asked the patients to rate their overall satisfaction with the hospital on a scale ranging from very poor to excellent. Each response from the patient will be correlated with overall satisfaction to determine which satisfaction dimension has the most vital relationship to overall satisfaction.

The questions in the questionnaire are related to the following dimensions: information, reliability, empathy, tangible, assurance, and responsiveness.

Information: All survey questions focus on patients' perceptions of information provided during their visit. This includes information related to delays and waiting times for their treatment. It also explains laboratory results, home-care instructions, information delivery in a language patients can easily understand, and the signage used in the healthcare facility.

Reliability: This factor addresses all issues related to waiting times during patients' healthcare facility visits. It also includes questions related to privacy and confidentiality.

Empathy: Questions address the healthcare team's considerations for the patient's respect and dignity.

Tangible: Questions related to the physical setting, such as the comfort of the waiting room, unit cleanliness, and availability of healthy food.

Assurance: These questions focus on the friendliness of the healthcare staff and whether consent is requested before diagnostic procedures are undertaken.

Responsiveness: Questions about patients' expectations and assessments of the healthcare team's willingness to answer questions from patients and their family members about how their pain is controlled.

The sample size is determined using Slovin's formula (Slovin, 1960), n=\(\:\frac{N}{(1+N{e}^{2})}\), where.

N = the size of the population.

E = margin of error

Using a 95% confidence interval in a sample size method (Nanjundeswaraswamy and Divakar, 2021), Slovin’s formula is used to compute the minimum number of necessary samples to meet the desired statistical constraints (Mgaya and Gwahula, 2024).

The content validity has also been assessed to confirm language clarity, practical relevance, and theoretical relevance in the Irish context (De Barros Aherns et al., 2020). Correspondingly, a panel of four members with recent hospital experience, two healthcare managers and two academics were asked whether the questionnaires measured the concept. Furthermore, experienced healthcare and academic teams are requested to assess face-to-face content validity to determine the appropriateness and relevance of the questions to the research questions. A pilot study is conducted over one week to ensure that the participants understand the questions clearly and quickly. The participants are asked to identify difficult questions.

The patient satisfaction questionnaire is mailed to the patients who attended the healthcare facility two weeks after they visited the Adult Emergency Department. The two weeks are selected to ensure that the participants memorise the details of their visit to the healthcare facility. Although using other methods of questionnaire delivery, such as phone surveys, might have a higher response rate, patients may have been more apt to report higher levels of satisfaction than using a more anonymous survey method, such as a mailed survey (Morgan et al., 2015). Furthermore, a mailed survey would encourage patients to rate their satisfaction openly and avoid possible threats to patient confidentiality and susceptibility to bias. Survey alert letters are sent to improve response rates and assure potential respondents that the research is legitimate and high-quality (Frandell et al., 2021). Postage-paid return envelopes are sent to randomly selected patients to encourage the participants to complete the questionnaires and avoid visiting the post office.

An information leaflet and consent form explaining the research purposes is mailed along with the questionnaire. The information leaflets contain a complete explanation of the study, including its aim, purpose, methods and procedures for data collection and the process for protecting their identity. Ethical approval was obtained from the ethics committee of the participating hospital before conducting the study.

Statistical analysis

Previous studies prove that Structured Equation modelling (SEM) is a crucial tool for managers and policymakers in healthcare (Avkiran, 2018). Among different SEM methods, partial least squares (PLS) were applied in this study. PLS is one of the SEM techniques that is well suited to assess complex predictive models. SEM is a common data analysis used in business research (Guenther et al., 2023). The two most common methods that dominate SEM are covariance-based SEM (CB-SEM) and partial least squares SEM (PLS-SEM). Whereas PLS-SEM focuses on explaining the variance in the model’s dependent variables, CB-SEM is mainly used to confirm or reject theories and their underlying hypotheses (Vinkóczi et al., 2024). PLS-SEM confirms/rejects hypotheses by determining how closely a proposed theoretical model can reproduce the covariance matrix for an observed sample dataset (Vaithilingam et al., 2024).

PLS is adopted to analyse the data using SMART-PLS software. Patient satisfaction determinants were evaluated by estimating multivariate models using PLS analysis. Path models are diagrams that visually display the hypotheses and variable relationships examined when SEM is applied. Model fit indices should be assessed before interpreting the PLS modelling results (West et al., 2023). The statistical analysis accordingly explores and presents the following:

Descriptive statistics for the patient’s characteristics.
Regression models to test hypotheses and explore the association between patient satisfaction and satisfaction dimensions.
Overall patients’ satisfaction as a dependent outcome measure.

Descriptive statistics were used to describe the demographic characteristics of the study sample. Table 7.1 Summarises the patients' clinical and demographic characteristics in the satisfaction questionnaire surveys. There was no predominance of male (45%) or female (55%) participants, indicating that the sample was not biased toward a particular gender. In this sample, 71% of participants were younger than 50 years old. Twenty-four patients visited the hospital more than once, while most of the patients visited the hospital at least once. The typical referral types include GPs (39%), self-referral (46%), ambulances (9%), and clinics (2%). Between noon and 6 pm, 36% of patients visited the ED. The percentage of participants admitted to the hospital at the end of their visit was 66.3%, whereas 32.5% of participants were discharged home.

The questions employ a five-point Likert scale ranging from very poor to excellent. The option of not applicable was added, as not all patients went through all the treatment stages, including diagnostic investigations and medication prescriptions. At the end of the questionnaire, the patients were asked to rate their overall satisfaction level.

Table 1

Clinical and demographic information for patients taking satisfaction survey.
Characteristic	Frequency	(%) of patients
Number	148
Gender
Female	81	55%
Male	67	45%
Age Group
< 18	03	02%
18–24	33	22%
25–34	34	34%
35–50	39	26%
51–65	21	14%
> 65	18	12%
Educational Level
Primary	13	09%
Secondary	64	43%
Third Level	55	37%
Referral Type
Ambulance	13	09%
Clinic	03	02%
G.P.	57	39%
Self-Referral	68	46%
Other	07	05%
Time of Arrival
Midnight – 8 a.m.	09	06%
8 a.m. – Noon	32	22%
Noon – 6 p.m.	54	36%
6 p.m. - Midnight	53	36%

The data reveal that 36.3% of the participants rate their satisfaction as fair or less, including poor and very poor, compared to 35% and 28.7% rate their satisfaction as very good and excellent, respectively (Fig. 1).

At the next stage, the reliability and validity must be verified by examining the explanatory and predictive power of the proposed patient satisfaction model. The conceptual PLS model presented in Fig. 2 was used. PLS-SEM involves a two-step approach, including estimating and solving the blocks of the measurement model and later analysing and estimating the path coefficients in the structural model (Schuberth et al., 2023). A SMART-PLS software was used to estimate the PLS model.

Measurement model

The measurement model (known as the outer model) represents the relationships among the constructs and the indicator variables. As per the PLS conceptual model (Fig. 2), six models are measured: tangibility, Information, assurance, empathy, responsiveness, and reliability. Tangibility was measured by 5 indicators, Information and reliability were measured by six indicators, and empathy and assurance were measured by 5 indicators. Content validity Individual item reliability, internal consistency and convergent validity are confirmed using a measurement model (Cheung et al., 2024). Reliability and validity are the two major methods for assessing the quality of the measurement model (Sureshchandar et al., 2023). Reliability is the degree to which data collection [tools and techniques] produces consistent results when the measured unit has not changed (Nawi et al., 2020). Furthermore, Convergent validity refers to the extent to which a test measures the same thing as other tests purported to measure that construct (Lim, 2024). Content validity was tested to confirm the language clarity, practical relevance, and theoretical relevance (Marie et al., 2021). Correspondingly, the same panel experts who were asked for their opinion about the patient’s expectations were inquired once more to review the patient’s questionnaires.

Furthermore, the experienced healthcare and academic teams were asked to conduct face-to-face content validity to identify the questions' appropriateness and relevance to the research questions. The questionnaires were revised based on the panel's feedback until it was judged to be ready. A pilot study was conducted over one week to ensure that the questions were comprehensible, clear and easily understood by the participants. The patient satisfaction questionnaires were distributed to 13 patients with a response rate of 84%. Participants were asked to provide feedback to identify difficult questions. There were no critical comments about any of the questions received from the respondents. Consequently, no changes were made to the questions. Additionally, cognitive interviewing was conducted with respondents to determine how they understood the questions and how they selected their responses.

The internal consistency reliability was assessed using Cronbach’s Alpha (CA) and composite reliability (CR) values, which offer powerful evidence of the measurement's reliability (Kalkbrenner et al., 2023).

Cronbach’s alpha was developed as a measure to assess the internal consistency of a set of scale or test items and is mostly used when the research has multiple-item measures of concept. According to (Hanafiah, 2020), CA and CR must be ≥ 0.7. Table 2 contains the convergent validity and reliability of the indicator variables. Composite reliability and the average variance extracted (AVE) are usually employed to access convergent validity (Hair et al., 2011). The strength of the relationship linking the patient’s satisfaction to the dimensions indicates the convergent validity of the construct. The average variance extracted (AVE) should be at least 0.50. As presented in Table (2), CA and CR are both more ≥ than 0.7, and the AVE value is more than 0.50. The results indicate that the model has satisfactory reliability and validity.

Multicollinearity is required to be tested next. Multicollinearity must be examined to confirm that there is no existing high correlation among two or more independent variables. Accordingly, exogenous latent constructs in PLS procedures are not supposed to be highly correlated (Hair and Sarstedt, 2019). Multicollinearity affects the statistical significance of the construct. Also, the variance inflation factor was measured. VIF was used in many studies as it is considered the reciprocal of the tolerance value. Accordingly, small VIF values indicate a low correlation among variables and vice-versa (Thoma et al., 2018). If VIF values are greater than 5, thus it represents high multicollinearity (Wondola et al., 2020). As demonstrated in Table 3, there is no high correlation between the latent constructs and VIF values of less than 5. Hence, there is no multicollinearity problem among latent variables.

Table 2

Convergent validity and reliability.
Constructs	Variables	CR	AVE	AVF
Tangibility (TNG)	T1	0.801	0.652	1.42
	T2
	T3
	T4
	T5
Information (INFO)	I1	0.864	0.729	2.95
	I2
	I3
	I4
	I5
	I6
Assurance (ASS)	A1	0.812	0.687	2.52
	A2
	A3
Empathy (EMP)	E1	0.855	0.714	1.98
	E2
	E3
Responsiveness (RES)	Res1	0.838	0.663	1.87
Responsiveness (RES)	Res2	0.838	0.663	1.87
Reliability (REL)	Rel1	0.883	0.743	2.21
	Rel2
	Rel3
	Rel4
	Rel5
	Rel6
Satisfaction		0.820	0.702

The significance and relevance of the indicators can be assessed through their outer weight using a bootstrapping procedure (Cheah et al., 2021). The bootstrapping procedure tests whether the outer weights are significantly different from zero. To assess significance, one can start bootstrapping with 10,000 sub-samples to check whether outer weights are significantly different from zero, as recommended by Hair and Sarstedt (2019).

Table 3

Coefficient correlation matrix and VIF.
Dimensions	TNG	INFO	ASS	EMP	RES	REL	Satisfaction	VIF
Tangibility	1.000	0.00	0.00	0.00	0.00	0.00	0.00	2.52
Information	0.422	1.000	0.00	0.00	0.00	0.00	0.00	2.35
Assurance	0.478	0.312	1.000	0.00	0.00	0.00	0.00	1.98
Empathy	0.325	0.462	0.299	1.000	0.00	0.00	0.00	2.67
Responsiveness	0.356	0.354	0.258	0.334	1.000	0.00	0.00	2.25
Reliability	0.481	0.282	0.385	0.291	0.387	1.000	0.00	2.34
Satisfaction	0.42	0.39	0.33	0.49	0.50	0.37	1.000	2.64

Also, guidelines were provided by (Legate et al., 2021) to manage the non-significant indicator weights. The researcher should consider either the absolute contribution [provided measuring the weight in PLS-SEM] or the absolute importance of the indicator to the construct. Accordingly, if the indicator outer weight is not significant (p > .05) but at the same time its outer loading is > 0.50, the research should decide either to retain it or remove it based on its absolute importance. On the other hand, if the indicator's outer weight is not significant (p > .05) and at the same time, its outer loading is < 0.5, and there is no evidence for its conceptual relevance, the research should remove the indicator from the measurement model. Table (4) shows all indicators are statistically significant, excluding 6. The non-significant indicators were retained related to their theoretical relevance, and their outer loading is > 0.5, as recommended by (Hair et al., 2011).

Structural Model

The structural model, known as the inner model, displays the relationship among the latent variables in the proposed model. After confirming that the outer model's variables are reliable and valid, the next stage is to assess the structural model (inner model). According to (Hair et al., 2014), five stages to be followed as the criteria for evaluating the structural model assessment procedure:

1) Asses whether the structural model has a collinearity issue or not. As per Table 3, Coefficient correlation matrix and VIF, there is no collinearity issue

2) Asses the significance of the path coefficient

3) Evaluate the coefficient of determination (R)

4) Evaluate the effect size (f)

5) Evaluate the predictive relevance (Q).

Assess the significance of the path coefficient

The next step in the structural model analysis is evaluating the significance level of the hypothesised relationships (i.e., path coefficients) among the constructs. Path coefficients represent the main outcomes of PLS-SEM, quantifying the structural model's hypothesised relationships (Kukah et al., 2024). In PLS, the tool for investigating the significance of path coefficients is the bootstrapping technique, which has a standardised value from − 1 to + 1 and is interpreted the same as standardised regression coefficients (Vishnoi et al., 2024). It tries to estimate the sampling distribution of a statistic by re-sampling with replacement from the original sample.

Table 4

Construct Outer Weights Significance Testing Results.
	Variable	Outer weight	Outer loading	T value	P value	Sig. (P < 0.05)
Tangibility (TNG)	T1	0.310	0.788	3.15	0.00	Yes
	T2	0.271	0.665	2.85	0.006	Yes
	T3	0.182	0.716	1.92	0.061	No
	T4	0.143	0.753	1.431	0.171	No
	T5	0.217	0.875	2.23	0.022	Yes
Information (INFO)	I1	0.263	0.723	2.71	0.007	Yes
	I2	0.225	0.846	2.28	0.044	Yes
	I3	0.103	0.698	1.2	0.209	No
	I4	0.259	0.733	2.621	0.008	Yes
	I5	0.152	0.765	2.163	0.128	No
	I6	0.218	0.872	2.25	0.021	Yes
Assurance (ASS)	A1	0.286	0.665	3.1	0.012	Yes
	A2	0.297	0.856	3.05	0.002	Yes
	A3	0.250	0.761	2.588	0.016	Yes
Empathy (EMP)	E1	0.090	0.752	0.19	0.243	No
	E2	0.223	0.851	2.245	0.027	Yes
	E3	0.257	0.739	2.581	0.016	Yes
Responsiveness (RES)	Res1	0.250	0.761	2.613	0.009	Yes
Responsiveness (RES)	Res2	0.217	0.875	2.24	0.024	Yes
Reliability (REL)	Rel1	0.127	0.825	1.26	0.205	No
	Rel2	0.132	0.728	1.35	0.201	No
	Rel3	0.275	0.691	2.81	0.007	Yes
	Rel4	0.215	0.883	2.252	0.019	Yes
	Rel5	0.264	0.721	3.12	0.001	Yes
	Rel6	0.222	0.854	2.81	0.006	Yes

The bootstrapping procedures using 5000 sub-samples were applied as recommended by (Hair et al., 2014) to be able to measure the significance of the path coefficients through t values and p values. As per the proposed construct within the patient satisfaction framework, there are 6 proposed relationships: tangibility H1, assurance H2, reliability H3, empathy H4, information H5 and responsiveness H6. Figure 3 presents the hypotheses' direction and the constructs' relationship. The six hypotheses are proposed in line with the literature review for the factors affecting the patient’s satisfaction. A hypothesis is a tentative statement about the relationship between two or more variables (Lawal et al., 2024).

Hypothesis 1

(H1): Tangibility positively influences patient satisfaction.

Hypothesis 2

(H2): Assurance positively influences patient satisfaction.

Hypothesis 3

(H3): Reliability positively influences patient satisfaction.

Hypothesis 4

(H4): Empathy positively influences patient satisfaction.

Hypothesis 5

(H5): Information positively influences patient satisfaction.

Hypothesis 6

(H6): Responsiveness positively influences patient satisfaction.

Hypothesis 7

(H7): Reliability mediates the relationship between the information and patient satisfaction.

Hypothesis 8

(H8): Assurance mediates the relationship between the information and patient satisfaction.

Hypothesis 9

(H9): Responsiveness mediates the relationship between reliability and patient satisfaction.

Hypotheses Results

The hypotheses results are presented in Table (5) below.

Hypothesis 1

(H1): Tangibility positively influences patient satisfaction.

Surprisingly, the direct and positive relationship between tangibility and patient satisfaction was not supported. The relationship structural path had a low coefficient (B = 0.184, P > 0.05), which was not statistically significant. Accordingly, the hypothesis was rejected.

Hypothesis 2

(H2): Assurance positively influences patient satisfaction.

The direct and positive relationship between assurance and patient satisfaction was supported. Path coefficient reveals that assurance is strongly associated with patient satisfaction with a significant value (β = 0.252, P < 0.05).

Hypothesis 3

(H3): Reliability positively influences patient satisfaction.

The path coefficients reveal that reliability is strongly associated with negative patient satisfaction (β =- 0.425, P < 0.05).

Hypothesis 4

(H4): Empathy positively influences patient satisfaction.

Surprisingly, the direct and positive relationship between assurance and patient satisfaction was not supported (β = 0.179, P > 0.05). Accordingly, the hypothesis was rejected.

Hypothesis 5

(H5): Information positively influences patient satisfaction.

The direct and positive relationship between assurance and patient satisfaction was supported. Path coefficient reveals that assurance is strongly associated with patient satisfaction with a significant value (β = 0.531, P < 0.05).

Hypothesis 6

(H6): Responsiveness positively influences patient satisfaction.

The direct and positive relationship between assurance and patient satisfaction was supported. Path coefficient reveals that assurance is strongly associated with patient satisfaction with significant value (β = 0.306, P < 0.05).

Further analysis of the model was conducted, and the mediation relationships were examined (Hypothesis 7, 8 and 9). The indirect path between information and reliability as a mediator is supported with significant value (β = 0.283, p < 0.05). Likewise, the indirect relationship between information and assurance is also supported by a significant value (β = 0.258, p < 0.05). This indicates that the information influenced patient satisfaction directly and indirectly through assurance and reliability.

The indirect relationship between reliability and responsiveness was not supported with a significant value (β = 0.167, P > 0.05). These indirect effects significantly result in understanding the relationships between the dimensions and their impact on patient satisfaction.

Table 5. Hypothesis test results.

Evaluate the Coefficient of Determination (R² Value)

The Coefficient of Determination (R²) demonstrate the extent of variability accounted for by the exogenous variable in its endogenous counterpart measures the model’s predictive power. The R² values range from 0 to 1, with higher levels indicating greater predictive accuracy and model fit. Chin (1998) considered the coefficient values of 0.67, 0.33, and 0.19 in PLS-SEM to be significant, moderate, and weak correspondingly.

Similarly, Sarstedt (2014) recommended R² values of 0.75, 0.50, and 0.25 and labelled them as substantial, moderate, or weak, respectively. In this study, the value R² is obtained as 0.792 (Table 6). According to the above recommendation, the model can be considered significant or substantial since it exceeded 0.67 and 0.75%. Accordingly, the dimensions have accounted for 79.2% of the variability in patient satisfaction.

Table 6

R² and Adjusted R² Values.
Latent Construct	R-Square	R-Square adjusted
Patient Satisfaction	0.792	0.785

Evaluate the Effect Size (f ²)

The next step was to measure the effect size (𝑓²). Measuring 𝑓² represents the changes in R² value for the endogenous construct if the exogenous construct is omitted in the model. The effect size values are 0.02, 0.15, and 0.35, representing small, medium, and large effects, respectively. The 𝑓² value below 0.02 is an indication of no effect. As presented in Table (7), the effect size for tangibility and empathy was small, medium for information and assurance, and large for reliability.

Table 7

f² Sizes Effect.
Latent Construct	F-Square	Effect Size
Tangibility	0.045	Small
Information	0.25	Medium
Assurance	0.18	Medium
Empathy	0.068	Small
Responsiveness	0.22	Medium
Reliability	0.36	Large

Evaluate the Predictive Relevance Q²

The next step in the analysis is to measure the Predictive Relevance [Q²] using Stone-Geisser’s Q2 value (Stone, 1974; Geisser, 1974). The Stone–Geisser criterion recommends that the model must be able to provide a prediction of the endogenous latent variable’s indicators (Vishnoi et al., 2024). If the Q² values are greater than zero, the path model has good predictive relevance. In PLS-SEM, the blindfolding procedure is used. The measured Q² is 0.328 (greater than zero), as presented in Table 8, and can be regarded as a medium. It suggested a good fit in model prediction.

Table 8

Q2 Predictive Relevance.
Latent Construct	SSO	SSE	Q² = 1-SSE/SSO
Tangibility	1350	1350
Information	2039	2039
Assurance	1890	1890
Empathy	2320	2320
Responsiveness	1560	1560
Reliability	1733	1733
Patient Satisfaction	1450	975	0.328

Comparing the path coefficient, it reveals and guides ED decision-makers to focus their attention on the information provided to the patients (β = 0.531), which is strongly associated with positive patient satisfaction with a medium-sized effect (f²=0.25) as per our results. Accordingly, adequate information about waiting time, diagnosis, medication and their side effect and discharge instructions must be understandable. The result is supported by the literature, which states that communicating with patients and their families has been reported as the strongest predictor of patient satisfaction. Moreover, reliability (β =-0.425) is strongly associated with negative patient satisfaction and requires ED decision-makers priority as it also measured the highest effect size (f²=0.36). Reliability focuses mainly on the patient's acceptable time to see the doctor, diagnostics, procedures, and transfer to the ward if the patient is admitted. The literature supports the result as it is considered a strong predictor of patient satisfaction. Waiting times have been widely publicised, negatively affecting patient safety and increasing morbidity (Seo et al., 2024). As per the literature, satisfaction related to reliability, especially long waiting hours, is affected by the information provided (Ye and Wu, 2024). The study supported the literature by examining the mediating effect of reliability on information and patient satisfaction (Hypothesis 7).

Unsurprisingly and like the findings of the literature (Mohammadi-Sardo and Salehi, 2019), the construct of responsiveness (β = 0.306) is associated positively with patient satisfaction as it is related to direct contact with the patients and their families, answering their questions and addressing the patient’s pain immediately. Willingness to help and answer patients' and their families’ questions is a gap indicated by the literature (Singh and Sharma, 2016) and an area that requires enhancement in healthcare. Many reasons contribute to the deficit in this area, including the workload in ED, which affects the healthcare team's efforts to build therapeutic relationships with the patients and their families (Kwame and Petrucka, 2021). Surprisingly, the mediating effect of responsiveness on reliability and patient satisfaction is not supported.

ED managers should consider the construct of assurance (β = 0.252) a further priority as it positively enhances patient satisfaction. Assurance is related to the healthcare team's knowledge, skills, and courtesy. The healthcare team plays a crucial role in the patient's feeling of assurance by obtaining the required knowledge and skills to treat the patient. The patient should believe that the healthcare team is medically competent.

Surprisingly, although some studies (Kalaja, 2023) reported tangible as an influential predictor for patient satisfaction, it is not positively associated with patient satisfaction (β = 0.184). In contrast to other studies (Handoko and Handayani, 2023), empathy (β = 0.179) was not also positively associated with patient satisfaction. More investigation in the future in this area is required to determine the main reasons. One potential reason is the patients' focus on getting the right treatment with less emphasis on the waiting room, the food or healthcare empathy.

Understanding the determinants of patient satisfaction allows policymakers and healthcare decision-makers to improve services in the Emergency department. Using the structural equation modelling approach, a patient satisfaction model was developed. According to our results, information, responsiveness, and assurance have proven to be positively associated with patient satisfaction in our model. In contrast, patient satisfaction was negatively influenced by reliability. Furthermore, only 63.7% of the participants reported their satisfaction as very good and excellent. Accordingly, more attention is needed on patient satisfaction, focusing on the determinant of patient satisfaction recognised in this study.

Author Contributions: Conceptualization, E.S.; Methods, E.S.; Formal Analysis, E.S. and W.R.; Writing—original draft preparation, E.S.; Writing— review and editing, W.R.; Supervision, W.R. and W.A.H.

Funding: This research received no external funding.

Institutional Review Board Statement: Before the study started, ethical approval was obtained from the healthcare facility's research ethics committee.

Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement: Not applicable.

Conflicts of Interest: The authors declare no conflicts of interest.

Alibrandi A, Gitto L, Limosani M, Mustica PF (2023) Patient satisfaction and quality of hospital care. Evaluation and Program Planning, 97, p.102251
Alowad A, Samaranayake P, Ahsan K, Alidrisi H, Karim A (2021) Enhancing patient flow in emergency department (ED) using lean strategies–an integrated voice of customer and voice of process perspective. Bus Process Manage J 27(1):75–105
Anderson TS, Thombley R, Dudley RA, Lin GA (2018) Trends in hospitalisation, readmission, and diagnostic testing of patients presenting to the emergency department with syncope. Ann Emerg Med 72(5):523–532
Atsavapranee E, Heidenreich P, Smith-Bentley M, Vyas A, Shieh L (2023) Care to share? Patients in private rooms are more likely to recommend a hospital to others—Journal. Patient Experience 10:23743735231158250
Avkiran NK (2018) An in-depth discussion and illustration of partial least squares structural equation modelling in health care. Health Care Manag Sci 21(3):401–408
Beecher C, Drummond L, Foley C, White M, Greene R, Ryan E, O’Dwyer LM, Dowling M, Devane D (2021) Development of a survey instrument to evaluate women’s experiences of their maternity care. Women Birth 34(4):e396–e405
Bhat S, Antony J, Gijo EV, Cudney EA (2020) Lean Six Sigma for the healthcare sector: a multiple case study analysis from the Indian context. Int J Qual Reliab Manage 37(1):90–111
Cheah JH, Roldán JL, Ciavolino E, Ting H, Ramayah T (2021) Sampling weight adjustments in partial least squares structural equation modelling: guidelines and illustrations. Total Qual Manage Bus Excellence 32(13–14):1594–1613
Cheung GW, Cooper-Thomas HD, Lau RS, Wang LC (2024) Reporting reliability, convergent and discriminant validity with structural equation modelling: A review and best-practice recommendations. Asia Pac J Manage 41(2):745–783
Chin WW (1998) The partial least squares approach to structural equation modelling. Mod methods Bus Res 295(2):295–336
De Ahrens B, da Silva Lirani R, L. and, de Francisco AC (2020) Construct validity and reliability of the work environment assessment instrument WE-10—. Int J Environ Res Public Health 17(20):7364
Department of Health (2008) Using the Commissioning for Quality and Innovation (CQUIN) payment framework. Department of Health, London, UK
Ferreira DC, Vieira I, Pedro MI, Caldas P, Varela M (2023) Patient Satisfaction with Healthcare Services and the Techniques Used for its Assessment: A Systematic Literature Review and a Bibliometric Analysis. Healthc (Basel) 11(5):639
Frandell A, Feeney MK, Johnson TP, Welch EW, Michalegko L, Jung H (2021) The effects of electronic alert letters for internet surveys of academic scientists. Scientometrics 126:7167–7181
Geisser S (1974) A predictive approach to the random effect model. Biometrika 61(1):101–107
Guenther P, Guenther M, Ringle CM, Zaefarian G, Cartwright S (2023) Improving PLS-SEM use for business marketing research. Ind Mark Manage 111:127–142
Hair JF, Christian M, Ringle, Marko S (2011) PLS-SEM: Indeed, a silver bullet. J Mark Theory Pract 19(2):139–152
Hair JF Jr, Sarstedt M (2019) Factors versus composites: Guidelines for choosing the right structural equation modelling method. Project Manage J 50(6):619–624
Hair JF, Hult GTM, Ringle CM, Sarstedt M (2014) A Primer on Partial Least Squares Structural Equation Modelling (PLS-SEM) (1 Ed). Sage, Thousand Oaks, CA
Hanafiah MH (2020) Formative vs. reflective measurement model: Guidelines for structural equation modelling research. Int J Anal Appl 18(5):876–889
Han S, Xu M, Lao J, Liang Z (2023) Collecting patient feedback as a means of monitoring patient experience and hospital service quality–learning from a government-led initiative. Patient Prefer Adherence, pp.385–400
Handoko N, Handayani SD (2023) Service quality and its effect on inpatient satisfaction at Purwodadi islamic hospital. Jurnal Aisyah: Jurnal Ilmu Kesehatan, 8(2)
Henstenburg JMD, Shah SAMD, Carrion, Rose BSH, Josephson GMD, MBA (2022) The Role of Satisfaction Surveys: Offering an Enhanced Patient Experience for Optimum Outcomes in the Pediatric Orthopedic Practice. J Pediatr Orthop 42:S13–S17
Hughes LD, Witham MD (2018) Causes and correlates of 30 day and 180 day readmission following discharge from a medicine for the elderly rehabilitation unit. BMC Geriatrics, 8; pp: 1–10
Kalaja R (2023) Determinants of patient satisfaction with health care: a literature review. Eur J Nat Sci Med 6(1):43–54
Kalkbrenner MT (2023) Alpha, omega, and H internal consistency reliability estimates: Reviewing these options and when to use them. Couns Outcome Res Evaluation 14(1):77–88
Kukah ASK, Owusu-Manu DG, Badu E, Asamoah E (2024) Structural equation model (SEM) for evaluating interrelationships among risks inherent in Ghanaian public-private partnership (PPP) power projects. Engineering. Constr Architectural Manage 31(6):2327–2352
Kwame A, Petrucka PM (2020) Communication in nurse-patient interaction in healthcare settings in sub-Saharan Africa: A scoping review—International. J Afr Nurs Sci 12:100198
Legate AE, Hair JF Jr, Chretien JL, Risher JJ (2023) PLS-SEM: Prediction‐oriented solutions for HRD researchers. Hum Res Dev Q 34(1):91–109
Lawal Y, Ayi OJ, Idoko AI (2024) A study of common errors in hypothesis formulation and testing among university students in social and management sciences. J Adv Res Multidisciplinary Stud 4(2):48–60
Lim WM (2024) A typology of validity: content, face, convergent, discriminant, nomological and predictive validity. J Trade Sci
Mgaya TW, Gwahula R (2024) The Effect of Service Quality on Customer Satisfaction at St. Joseph Referral Hospital-Peramiho, Tanzania. Eur J Theoretical Appl Sci 2(2):223–233
Mohammadi-Sardo MR, Salehi S (2019) Emergency department patient satisfaction assessment using modified servqual model; a cross-sectional study. Advanced Journal of Emergency Medicine, 3, (1)
Morgan MW, Salzman JG, LeFevere RC, Thomas AJ, Isenberger KM (2015) Demographic, operational, and healthcare utilisation factors associated with emergency department patient satisfaction. Western J Emerg Med 16(4):516
Nawi FAM, Tambi AMA, Samat MF, Mustapha WMW (2020) A review on the internal consistency of a scale: the empirical example of the influence of human capital investment on Malcom Baldridge quality principles in TVET institutions. Asian People J (APJ) 3(1):19–29
Nanjundeswaraswamy TS, Divakar S (2021) Determination of sample size and sampling methods in applied research. Proceedings on Engineering Sciences, 3(1), pp.25–32
Novitasari D (2022) Hospital Quality Service and Patient Satisfaction: How is the Role of Service Excellent and Service Quality? J Inform Syst Manage (JISMA) 1(1):29–36
Omaghomi TT, Akomolafe O, Onwumere C, Odilibe IP, Elufioye OA (2024) Patient experience and satisfaction in healthcare: a focus on managerial approaches-a review. Int Med Sci Res J 4(2):194–209
Okolo CA, Ijeh S, Arowoogun JO, Adeniyi AO, Omotayo O (2024) Healthcare managements role in addressing disparities: a review of strategies. Int J Appl Res Social Sci 6(4):518–531
Park SY, Yun GW, Friedman S, Hill K, Coppes MJ (2022) Patient-centred care and healthcare consumerism in online healthcare service advertisements: a positioning analysis. Journal of Patient Experience, 9, pp.23743735221133636
Poorani AA, Kline SF, DeMicco FJ, Sullivan W (2023) Hospitality to healthcare: Patient Experience Academy, a successful alliance between the ChristianaCare Health System and the University of Delaware. International Journal of Hospitality Management, 112, pp.103422
Saul E (2024) Visual Identity Guidelines For HSE and Funded Agencies
Schuberth F, Rademaker ME, Henseler J (2023) Assessing the overall fit of composite models estimated by partial least squares path modeling. Eur J Mark 57(6):1678–1702
Seo H, Ahn I, Gwon H, Kang HJ, Kim Y, Cho HN, Choi H, Kim M, Han J, Kee G, Park S (2024) Prediction of hospitalisation and waiting time within 24 hours of emergency department patients with unstructured text data. Health Care Manag Sci 27(1):114–129
Sheard L, Peacock R, Marsh C, Lawton R (2019) What's the problem with patient experience feedback? A macro and micro understanding, based on findings from a three-site UK qualitative study. Health Expect 22(1):46–53
Singh S, Sharma Y (2016) Evaluation of quality in hospitals of Haryana: A perspective of doctors and nurses. Int J Curr Eng Technol 6(4):1422–1430
Sonis JD, Aaronson EL, Castagna A, White B (2019) A conceptual model for emergency department patient experience. J Patient Experience 6(3):173–178
Sureshchandar GS (2023) Quality 4.0–a measurement model using the confirmatory factor analysis (CFA) approach. Int J Qual Reliab Manage 40(1):280–303
Marie S, Jimmerson B, Perkhounkova A, Y. and, Herr K (2021) Developing and establishing content validity of vignettes for health care education and research. West J Nurs Res 43(7):677–685
Stone M (1974) Cross-Validatory Choice and Assessment of Statistical Predictions. J Roy Stat Soc 36:111–147
Swallmeh E, Byers V, Arisha A (2018) Informing quality in emergency care: understanding patient experiences. Int J Health Care Qual Assur 1331(7):704–717
Thoma RJ, Cook JA, McGrew C, King JH, Pulsipher DT, Yeo RA, Campbell RA (2018) Convergent and discriminant validity of the ImPACT with traditional neuropsychological measures. Cogent Psychol 5(1):1430199
Vaithilingam S, Ong CS, Moisescu OI, Nair MS (2024) Robustness checks in PLS-SEM: A review of recent practices and recommendations for future applications in business research. Journal of Business Research, 173, p.114465
Vinkóczi T, Rácz ÉH, Koltai JP (2024) Exploratory analysis of zero waste theory to examine consumer perceptions of sustainability: A covariance-based structural equation modeling (CB-SEM). Cleaner Waste Systems, 8, p.100146
Vishnoi SK, Mathur S, Bagga T, Singhal A, Rawal P, Sharma S, Yadav R (2024) Construct modelling, statistical analysis and empirical validation using PLS-SEM: a step-by-step guide of the analysis procedure. Int J Data Anal Techniques Strategies 16(2):162–180
West SG, Wu W, McNeish D, Savord A (2023) Model fit in structural equation modeling. Handb Struct equation Model 2:184–205
Wondola DW, Aulele SN, Lembang FK (2020) Partial least square (PLS) method of addressing multicollinearity problems in multiple linear regressions (case studies: cost of electricity bills and factors affecting it). In Journal of Physics: Conference Series, 1463(1) p. 012006)
Ye Q, Wu H (2024) Understanding the role of information transparency in improving patient experience under different uncertainties: a quasi-natural experiment. Internet Res 34(3):718–747

The authors declare no competing interests.

Exploring the Determinants of Patient Satisfaction in Emergency Healthcare: Evidence from Irish EDs

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Background

Materials and Methods

Study Setting

Survey Instrument

Statistical analysis

Results

Measurement model

Structural Model

Assess the significance of the path coefficient

Hypotheses Results

Evaluate the Coefficient of Determination (R² Value)

Evaluate the Effect Size (f ²)

Evaluate the Predictive Relevance Q²

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1