Introduction
Schizophrenia affects nearly 1% of the world population (Marder & Cannon, 2019), and patients with schizophrenia also often have a higher prevalence of comorbidities such as cardiovascular disease, obesity, diabetes, dental problems, and skin diseases than the general population (Leucht et al., 2007).
After discharge, effective disease management programs play a vital role in schizophrenia management (Silva et al., 2009). These disease management programs are usually developed based on a chronic care model (Chen et al., 2021; Wagner et al., 2001) that includes characteristics of patient-centered care (PCC; Wagner et al., 2001) that may improve patient outcomes across different disciplines (Rathert et al., 2013; Tsvitman et al., 2021). In 2000, Mead and Bower conducted a review of the concept of patient-centeredness (Langberg et al., 2019). In 2001, the international Picker Institute concluded that PCC involves eight domains of care, including (a) respect for patient preferences, values, and demands; (b) collaboration and integration of healthcare services; (c) information, education, and communication; (d) emotional support; (e) physical comfort; (f) medical involvement of relatives and close friends; (g) care continuity and transformation (i.e., from hospital to family); and (h) convenient use of healthcare and services (Gerteis et al., 1993; Rathert et al., 2013, 2015).
In nursing care, implementing activities to enhance the self-efficacy, coping, and self-management skills of patients with mental illness to improve outcomes is an important task (Chan, 2021). In general, evidence-based practices (i.e., pharmacological interventions) combined with nonpharmacological treatments such as PCC-related activities enhance patients' involvement and self-efficacy and improve outcomes in patients with schizophrenia (Correll, 2020; Engle et al., 2021). Previous research related to schizophrenia care has shown that disease management programs with characteristics such as goal setting (Oles et al., 2015) and emotional support (Allerby et al., 2020) can help achieve better outcomes in terms of patient activation and satisfaction. However, studies designed to comprehensively evaluate the eight PCC factors and identify priority factors affecting outcomes in patients with schizophrenia are lacking. Therefore, this study was developed to examine the PCC factor models and determine the key factors that positively affect patient outcomes and, based on the findings, to suggest priority practical activities to improve patient performance in the related domains.
Methods
Patient Selection
The data for this descriptive, correlational survey study were collected from patient surveys and record reviews from two hospitals in northern Taiwan, including one general regional hospital and one psychiatric hospital. Trained nurses and a clinical psychologist distributed questionnaires to patients with schizophrenia between November and December 2016, and the participants were selected from a convenience sample of patients with schizophrenia. The minimum sample size of 120 was calculated based on the rule of thumb for regression analysis that specifies at least 10 cases be included per estimated parameter (Waltz et al., 2017) and on the 12 confounding factors addressed in the considered regression models. The inclusion criteria were patients (a) with a diagnosis of schizophrenia (ICD-9-CM 295), (b) 18 years or older, and (c) with sufficient mental capacity to respond to the complete questionnaire. In this study, 150 questionnaires were distributed, and 150 valid questionnaires were returned (return rate: 100%). The final analytical sample consisted of 150 individuals.
Data Collection Process
This research was approved by the ethics committee of the institutional review board, Cardinal Tien Hospital (IRB Number CTH-104-3-1-051). An initial pilot test was conducted on 10 patients with schizophrenia to verify that the format and questionnaire items were sufficiently clear for respondents. To ensure consistency among the interviewers in this study, interviewer training meetings were conducted at the two target hospitals before initiating the interview process. During each training meeting, each item and answer category were read aloud, and the purpose of each question was explained, particularly in terms of the actual differences among the questionnaire constructs. Next, any remaining questions from the interviewers were addressed. We confirmed that trained interviewers were able to read the questionnaire items fluently and follow the question wording exactly. In addition, the trained interviewers used standardized data collection forms to gather information from the treatment records of each patient. If a respondent could not read or was unfamiliar with the questionnaire, the interviewer read the entire questionnaire to the interviewee to ensure their complete understanding. After the participants had completed the questionnaire, the interviewers verified the accuracy of the demographic information provided by the patients.
Patient-Centered Care Measures
Three experts ensured the content validity of the questionnaire by verifying that the design was robust and that the questions were sufficiently clear for the targeted respondent group. Two concepts, namely, "patient-perceived physician support of their autonomy" and "goal setting," relate to patient preferences, values, and demands (Jaensch et al., 2019; Luther et al., 2019; Schmittdiel et al., 2008). The former refers to the support provided by healthcare professionals to patients that enables them to manage their chronic conditions well (A. A. Lee et al., 2019). The questionnaire adopted in this research study to assess this concept was the Health Care Climate Questionnaire (Williams et al., 2005), which was modified by using only six of the original 15 items published in a previous study (Y. Y. Lee & Lin, 2010). A sample item from this questionnaire is "I feel that my doctor provided me with choices and options for handling my schizophrenia." Reliability analyses indicated medium-high internal consistency (Cronbach's alpha = .70). The second concept, goal setting, refers to the act of helping patients with schizophrenia identify what they want to achieve, how they can achieve it, and where they can direct their efforts (Med-IQ, 2014). The questionnaire adopted in this study to assess this concept was the Patient Assessment of Chronic Illness Care (Schmittdiel et al., 2008), which includes one construct (five items), that is, goal setting/tailoring, as the goal-setting measure. A sample item from this questionnaire is "I am asked to talk about my goals in caring for my condition." Reliability analyses indicated high internal consistency (Cronbach's alpha = .78). The Patient Assessment of Chronic Illness Care questionnaire was also adopted in this study to assess the domain of collaboration and integration of healthcare services. A sample item related to this domain is "I am encouraged to attend programs in the community that can help me" (Gerteis et al., 1993; Rathert et al., 2015). The reliability analyses indicated high internal consistency (Cronbach's alpha = .86). For the domain of information, education, and communication, the Consumer Assessment of Healthcare Providers and Systems (Navarro et al., 2021) was adopted in this study. The original questionnaire includes six items related to communication, and Ratanawongsa et al. shortened the communication subdomain to include only four items (e.g., "Did doctors or healthcare providers spend enough time with you?"; Ratanawongsa et al., 2013). In this study, after eliminating one item (i.e., Did doctors or healthcare providers listen carefully to you?) from the four items because it overlapped with the first domain (support of autonomy), a three-item domain related to information, education, and communication was formed. A reliability analysis indicated that withdrawing one item was appropriate (e.g., the patient received answers he or she could understand from the medical team), as that item may present difficulties for patients with schizophrenia with cognitive deficits to understand. The Cronbach's alpha for the reliability analysis of this subdomain was .84. For the domain of emotional support, the Picker PCC questionnaire (Gerteis et al., 1993; Rathert et al., 2015) was adopted. A sample item related to this domain is "A nurse discussed anxieties and fears with me." The reliability analyses indicated high internal consistency (Cronbach's alpha = .78). All of the items asking about physicians or nurses only were rephrased to refer to medical professionals in general because disease management programs are team-based plans that require cooperation among professionals in different disciplines.
The PCC activities and their associated domains at the two hospital sites are shown in Table 1.
Outcome Measures
In PCC-related studies on schizophrenia care, concepts such as patient activation and patient empowerment represent primary outcomes only (Allerby et al., 2020; Oles et al., 2015), and Allerby et al.'s study suggested that patient satisfaction be adopted as a secondary/final outcome (Allerby et al., 2020). In addition, a more patient-centered model that evolved from a disease-centered model focuses on the patient's perspective on the disease and treatment in addition to the effects and side effects of treatment. The patient perspective is typically systematically measured in terms of patient-reported outcome measures (PROMs). In chronic disease research, one of the most commonly used PROMs is patient satisfaction (Cella et al., 2015). PROMs are important for patients with schizophrenia because pharmacologic interventions should be combined with evidence-based nonpharmacologic treatments and patient-reported outcomes should be employed (Correll, 2020). Finally, a systematic review study also highlighted that higher satisfaction may ultimately improve long-term, simple objective outcomes in patients (Navarro et al., 2021). Thus, based on the above, satisfaction was considered a final patient outcome in this study.
A standard generic-type satisfaction questionnaire, the Patient Satisfaction Questionnaire 18, was also used in this study (Ong et al., 2000), with four items related to waiting time and financial issues (e.g., I am usually kept waiting for a long time when I am at the doctor's office) removed to form a 14-item domain related to patient satisfaction. A sample item used in this questionnaire is "I am very satisfied with the medical care I receive." Reliability analyses indicated high internal consistency (Cronbach's alpha = .86).
Confounding Factors
In this study, patient characteristics were divided into demographic and clinical factors. Demographic factors included age, gender, education, occupation, marriage, and urbanization of residence area. Regarding urbanization of residence area, communities were stratified in this study based on the practice of the Taiwanese National Health Research Institute into seven level-of-urbanization categories, including (from highest to lowest) high-level, median-level, emerging, common, aging, agricultural, and remote areas (W. P. Chang et al., 2013). The clinical factors included the Clinical Global Impressions (CGI) severity (S) and improvement (I) indexes and previous admission, emergency department visit, or readmission within 1 year. Because of the few (2) participants in "markedly ill" category, the CGI-S scores for the participants in this category were combined with those of the participants in the moderately ill category for regression analysis.
Analysis
To adjust the nonnormal conditions for patient satisfaction scores after conducting a Shapiro-Wilk test of normality, we transformed the scores into 1.5 squared. We adopted two models to analyze the data. First, a multivariable linear regression based on ordinary least squares with stepwise selection and including all PCC factors and all confounders in one model was applied to all regressions. To be conservative, the significance levels for entry and stay were set to .15. The stepwise variable selection procedure (with iterations between the forward and backward steps) was helpful to obtaining the best candidate final regression model. The best candidate final regression model was identified manually by dropping the covariates with a p value > .05 one at a time until all of the regression coefficients differed significantly from 0. The goodness-of-fit measures included the coefficient of determination R2 (for the linear regression model). Second, to examine the relationships between the PCC environment at two hospitals and patient satisfaction, a generalized estimating equation (GEE) for the nested data structure was adopted, with patients nested in their associated hospitals. SAS software Version 9.4 (SAS Institute, Inc., Cary, NC, USA) was used for the above analysis.
Common method variance (CMV) introduces variance in scores attributable to the method used to measure a construct rather than the construct it represents (McDermott & Sharma, 2017). CMV may significantly inflate the associations between different constructs. As controlling for CMV using different time points in this study was difficult (i.e., one patient completing the same questionnaire at two different times), this study adopted several methods to prevent CMV bias (Podsakoff et al., 2003), including psychological isolation (e.g., we asked the interviewers to emphasize that every domain was independent before conducting interviews), randomization of item allocations (e.g., randomized mixture of items in different constructs), and reverse-item methods.
Results
Demographics of Patients With Schizophrenia
One hundred fifty patients with schizophrenia were interviewed. As shown in Table 2, the average age of the patients was 48 years, half of the patients were female (54%), half (47%) had a senior high school degree as their highest level of education, only 21% had a full-time or part-time job or were retired, 69% were unmarried, 47% lived in towns with a high level of urbanization, 64% had no comorbidities, 30% were borderline mentally ill (CGI-S score of 2), and 42% were much improved (CGI-I score of 2).
All Patient-Centered Care Factors for Satisfaction (Linear Regression)
As shown in Table 3, after controlling for all of the confounding factors listed in the methodology section, three PCC factors, including support of autonomy (parameter = 0.59 [0.18, 0.99], p = .005); information, education, and communication (parameter = 0.43 [0.03, 0.83], p = .036); and emotional support (parameter = 0.46 [0.15, 0.77], p = .004), showed significantly positive associations with patient satisfaction. Confounder patients with much improved conditions (CGI-I score of 2) were also associated with higher patient satisfaction (parameter = 0.50 [0.14, 0.86], p = .007) than patients with no change or who were minimally worse (CGI-I score of 4 or 5). The R2 value for satisfaction was .43.
All Patient-Centered Care Factors for Satisfaction (Generalized Estimating Equation Model)
As shown in Table 4, after controlling for all of the confounding factors listed in the methodology section, three PCC factors differed slightly from the results of the linear regression, including goal setting (parameter = 0.31 [0.10, 0.51], p = .004); information, education, and communication (parameter = 0.65 [0.37, 0.92], p < .001); and emotional support (parameter = 0.52 [0.22, 0.81], p < .001), each of which showed significantly positive associations with patient satisfaction. The confounders included patients who completed high school being associated with lower levels of satisfaction (parameter = -0.28 [-0.51, -0.06], p = .014) than patients with education below an elementary school, patients living in an area with emerging urbanization being associated with lower levels of satisfaction (parameter = -0.43 [-0.55, -0.31], p < .001) than patients living in areas with a high level of urbanization, and patients with conditions that made them moderately or markedly ill (CGI-S score of 4 or 5) being associated with lower levels of satisfaction (parameter = -0.50 [-0.81, -0.20], p = .001) than patients with a borderline mental illness (CGI-S score of 2).
Discussion
The authors' review of the related literature indicates that this study was the first conducted in recent years to evaluate satisfaction in patients with schizophrenia using a comprehensive PCC. The results showed that, after all PCC factors were included in the regression analysis, the three most important factors were (in descending order) as follows: information, education, communication and emotional support, and goal setting.
Our GEE model results showed that patients with a high level of education, low socioeconomic status (living in areas with low levels of urbanization), or high severity of disease (high CGI-S score) were associated with low satisfaction. In previous PCC-related studies on schizophrenia, being Black (low socioeconomic status) or highly educated was found to be associated with low patient activation, which is an indicator of low satisfaction (Mosen et al., 2007). In several PCC-related studies on cancer care, highly educated patients or patients with a high severity of disease (e.g., late stage of cancer) were found to exhibit low satisfaction (Robinson et al., 2013). All of the abovementioned evidence suggests the results of our study are valid.
Increased implementation of these three PCC factors was found to be associated with better patient outcomes, especially in terms of patient satisfaction. This finding was also highlighted in other studies on schizophrenia and other psychological conditions. The most important PCC factor for schizophrenia relates to information, education, and communication. A recent systematic review showed patient-provider communication to be significantly associated with self-reported physical and mental health (Navarro et al., 2021). Pestana-Santos et al. (2018) found that patients with schizophrenia consider clinical communication necessary when assessing psychiatrists' communication skills. Some studies addressing conditions other than schizophrenia have found that good communication by medical teams encourages patient adherence to medication use for diabetes care and is associated with better patient satisfaction (Consumer Assessment of Healthcare Providers and Systems score; T. J. Chang et al., 2021; Ratanawongsa et al., 2013). The second most important PCC factor for schizophrenia relates to emotional support. Incorporating emotions into clinical decision making is important (Zhang & Liao, 2021). Allerby et al. (2020) found a relationship between satisfaction and establishing partnerships between patients and medical teams using patient narratives to support active engagement in psychosis care. As an example of a study not related to mental health, Meterko et al. (2010) showed that good emotional support can produce patient trust, which may promote patient adherence to prescribed medication use, increase patients' likelihood of receiving examinations, and improve outcomes in patients with acute myocardial infarction. Regarding goal setting, review studies have shown motivational interviews for schizophrenia care addressing topics such as goal setting to be positively associated with patient-reported outcomes (Correll, 2020). Oles et al. (2015) found that goal setting can help foster feelings of hope in patients and thus improve outcomes for patients with schizophrenia. Other studies not related to schizophrenia also found that a patient-led goal-setting intervention can improve outcomes for chronic low back pain (Gardner et al., 2019).
Under the GEE model, professional coordination and support of autonomy were not shown to relate with patient satisfaction when all five PCC factors were considered together. However, this does not mean that they are not essential PCC factors. Regarding the first factor, although studies in other fields have shown that integrated care can encourage patients to use more preventive services and is associated with better glycated hemoglobin control in patients with diabetes (Aung et al., 2015), this was not the case in this study. This may be because of other factors being related to other treatment outcomes such as adequate functionality (Correll, 2020) and less related to positive patient satisfaction (delight) than other PCC factors of interpersonal relationships such as emotional support, communication, support of autonomy, and help with goal setting. Furthermore, schizophrenia is associated with distinctive disease characteristics such as lack of a sense of the disease and poor appreciation of treatment activities such as collaborative care. Regarding the last factor, studies on conditions other than schizophrenia have found support of patient autonomy to be positively associated with satisfaction and better outcomes (Jaensch et al., 2019). However, a recent review showed that, for schizophrenia care, a combined pharmacological intervention and shared decision making (SDM) are positively associated with patient-reported outcomes (Correll, 2020). Regarding patient preferences, values, and demands, in addition to support of autonomy and goal setting (Jaensch et al., 2019; Luther et al., 2019; Schmittdiel et al., 2008), several studies have addressed SDM as a relevant concept (Stiggelbout et al., 2015). Patients involved in the SDM process to identify their own treatment preferences and informed of the advantages and disadvantages of treatment alternatives may feel higher satisfaction than just through the perception of support of autonomy, which only emphasizes enabling patients to become good managers of their chronic conditions (A. A. Lee et al., 2019) and does not provide practical solutions. These explanations should be further investigated using in-depth interviews.
The clinical implications for nursing care for schizophrenia are that simple emotional support and good communication may enhance patient satisfaction. Hence, for nursing care, comforting patients and their relatives experiencing anxiety is the priority. Second, in addition to providing education about social interaction skills and information on the use of medications, clear and effective mutual communication to help address living conditions tailored to patients and their health literacy is also important. Finally, if possible, using motivational interviews such as goal setting to implicitly incentivize patients to participate in the treatment process is also of value in improving satisfaction with treatment. In the future, to further promote PCC, efforts should be made at an organizational level (Johnsen et al., 2022), and PCC content should be included in medical/nursing curricula (Engle et al., 2021) and incorporated in clinical guidelines (Kim et al., 2021).
There are several limitations to this research. First, the research included only five of the eight domains proposed by the Picker Institute. Some of the domains, such as physical comfort, medical involvement of family, care transformation (i.e., from inpatient to family), and convenient use of healthcare and services are not applicable for patients with schizophrenia. For example, physical comfort, care transformation, and convenient use of healthcare/services may not be significant concerns for patients with schizophrenia in outpatient settings under universal-coverage-based health insurance systems. The domain related to the medical involvement of families was also not assessed. This domain may not be directly perceived as a source of patient satisfaction, as the original questionnaire item (i.e., "medical teams gave family all the information they needed") may not directly relate to patient-perceived satisfaction. Future studies should further investigate the relationship between family involvement and patient satisfaction or other outcomes. Second, the research results should be interpreted conservatively and applied cautiously to other disciplines because of the small sample size used.
Conclusions
The three critical PCC-related factors proposed by the Picker Institute were selected to enhance patient satisfaction with their schizophrenia care program. Elements of these factors, including simple emotional support, good communication, education, and goal setting, all enhance patient satisfaction. Practicable activities related to the three PCC factors are also suggested in this article for implementation in the clinical setting.
Acknowledgments
This work was supported by the National Science and Technology Council (NSTC 105-2410-H-030-057) and the Cardinal Tien Hospital (CTH-103-1-2B05) in Taiwan. We are grateful to Miss Ya-Yuan Hsu for her assistance.
Author Contributions
Study conception and design: TTC, JJY, KHC
Data collection: TTC, JJY, KCC, CLC
Data analysis and interpretation: TTC
Drafting of the article: TTC, KCC, CLC
Critical revision of the article: JJY, KHC
References