Using Linear Regression to Identify Critical Demographic Variables Affecting Patient Safety Culture From Viewpoints of Physicians and Nurses

1Department of Business Administration, National Changhua University of Education, Changhua City, Taiwan 2School of Business Administration, Hubei University of Economics, Wuhan City, People’s Republic of China 3Department of Health Business Administration, Hung Kuang University, Taichung City, Taiwan 4School of Health Policy and Management, Chung Shan Medical University, Taichung City, Taiwan 5Quality Management Center, Shanghai Changtai Medical Technology Co. Ltd., Shanghai, People’s Republic of China


Background
The issues of patient safety and healthcare quality have become increasingly important across the world since the 1990s.Many hospitals manage to reduce the occurrence of adverse events (AEs) which are common in healthcare organizations.AEs due to a defective system or human negligence are enormous threats to patient safety.Previous studies in several countries have shown that approximately half of all AEs are considered preventable. 1Some healthcare initiatives or policies have been implemented to enhance the quality of healthcare.The National Patient that promote safety." 6 International institutions such as the Institute of Medicine (IOM) and the Joint Commission are encouraging healthcare organizations to assess patient safety and reinforce medical quality through safety culture surveys.Assessing the existing patient safety culture allows management to have a clear vision of an organization's strengths and weaknesses.
Various measures have been created for gauging a patient safety culture.8][9] The SAQ was derived from the Flight Management Attitudes Questionnaire and retained 25% of it.
0][11][12][13][14][15] The Joint Commission of Taiwan (JCT) developed the Chinese version of the SAQ (SAQ-C) using forward and backward translation.The intelligibility and item applicability of the questionnaire was confirmed by an expert panel. 15JCT modified the original questionnaire from six dimensions and 30 items to nine dimensions and 41 items.The 3 added dimensions (hospital management support for patient safety, teamwork across hospital units, and hospital handoffs and transitions) were integrated into the questionnaire.The SAQ-C current in 2014 retained 6 dimensions of the original SAQ and combined 2 new aspects, "emotional exhaustion" and "work-life balance", with nine and seven questions, respectively.The latest Chinese version is provided in Table 1. 16elevant research revealed that having different demographic information will lead to different perceptions of patient safety among healthcare providers. 2,17,18A study by Kim et al verified that nurses on the front line have more problems with patient safety than those who are older or who work in management positions. 19Physicians and nurses form the core staff of each healthcare organization. 17

Objective
They are the front line healthcare providers and have contact with patients at a high frequency.It would be of interest to observe how physicians and nurses with different demographic information perceive the patient safety culture.Therefore, this study aimed to identify the crucial demographic variables viewed by physicians and nurses as significantly influencing the patient safety culture in a regional teaching hospital in Taiwan.

Methods
This research was conducted in 2015 in a regional teaching hospital located in Taichung City, Taiwan.The studied hospital has all the major medical specialties and services and can admit approximately 700 patients.All the physicians and nurses in this hospital were invited to take part in the current study.After removing the invalid parentheses, the valid number of participants in this study was 376, comprising 42 physicians and 334 nurses.Data was collected internally in this case hospital through the 2014 SAQ-C from JCT which included 46 items that examine medical staff members' attitudes toward eight dimensions, i.e. teamwork climate, safety climate, job satisfaction, stress recognition, perceptions of management, working conditions, emotional exhaustion, and work-life balance.A 5-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree was used for responses for all dimensions except "work-life balance, " which used a 4-point scale to measure the frequency per week.Because of its different scale, the analysis of "work-life balance" was excluded from this study.Unlike the other professional groups, physicians and nurses were asked to answer all question items.The demographics of gender, age, supervisor/manager, job position, job status, experience in organization, experience in position, education, and direct patient contact were taken into account for analyses.
Negatively-worded items (Items 2, 11, 20, 22, 23, 24, 35,  36, 37, 38, and 39) in this questionnaire were reverse scored.Individual questionnaire responses were aggregated by computing the score of the items for each dimension.The demographic variables were independent variables, while the score of each dimension was considered a dependent variable.To predict the dependent variables from predictor variables, this study employed linear regression with forward selection, which started with an empty set and continually added one attribute at a time.At each step, only the attribute which gave the highest performance was added into the selection.The model selection criteria of this technique were tractable and disclosed which independent variable(s) (demographic variable(s)) were critical to a particular dimension. 20All statistical analyses were carried out using SPSS software version 18.The results of the linear regression with forward selection are presented in Table 3 with α = 0.05.The adjusted R-square values range from 0.048 to 0.138.Teamwork climate was negatively influenced by "supervisor/manager" and "experience in position, " indicating that physicians and nurses who are not in charge are less satisfied, and physicians and nurses who have much more experience in their positions tend to be less satisfied as well."Supervisor/ manager" and "experience in position" had negatively impacts on the safety climate.That is, physicians and nurses who are not in charge and who have more experience in their positions are less satisfied.
"Supervisor/manager", "age", and "experience in position" are the 3 critical demographic variables affecting job satisfaction.Physicians and nurses who are in charge, elderly, or less experienced in their positions tend to have higher job satisfaction.Stress recognition is impacted by job position and supervisor/manager.Specifically, nurses tend to have less satisfaction in stress recognition.
"Supervisor/manager" and "experience in position" are the 2 essential demographic variables that negatively influence perceptions of management.Physicians and nurses who are not supervisors/managers and/or have much experience in their positions tend to have less satisfaction in perceptions of management.The dimension of "working conditions" was impacted by "supervisor/ manager", "experience in organization", and "age".Physicians and nurses who have much experience in the organization feel less satisfied.In contrast, employees who are elderly are more satisfied with their working conditions.Finally, the linear regression model could not be established between the dimension of "emotional exhaustion" and ten demographic variables.By further examining the correlation between "emotional exhaustion" and the demographic variables, the P values ranging from 0.067 to 0.976 indicate that there is no correlation between them if α is set to 0.05.
The correlation of seven dimensions and critical demographic variables is shown in Table 4.In summary, physicians and nurses who are not supervisors/managers feel less satisfied with the teamwork climate, safety climate, job satisfaction, stress recognition, perceptions of management, and working conditions.Employees who have much experience in their positions tend to be less satisfied with the teamwork climate, safety climate, job satisfaction, and perceptions of management.In addition, employees who have much experience in an organization are less satisfied with the working conditions.On the other hand, physicians have less stress than nurses.Finally, older employees tend to have a high satisfaction with their jobs and working conditions.
The coefficients of most explanatory variables were negative, clarifying that the respondents who were supervisors/managers and who had much experience in their positions had the greatest effect on the patient safety culture in this investigation.Healthcare workers who were supervisors/managers in this hospital indicated a higher level of satisfaction on the SAQ-C with the teamwork climate, safety climate, job satisfaction, stress recognition, perceptions of management, and working conditions.Healthcare workers with less experience in their positions in this hospital were inclined to have better perceptions of the teamwork climate, safety climate, job satisfaction, and perceptions of management.Nevertheless, "age" was the only positive demographic variable.It is worth mentioning that job position had some bearing on patient safety culture.

Discussion
This study utilized the SAQ-C to assess the patient safety culture and identify the crucial demographic variables from the viewpoints of 376 physicians and nurses in 2015.Linear regression with forward selection was applied to analyze the data; results indicated that the predictor variable "supervisors/managers" was the most important variable for the respondents in this research.
Employees who serve as supervisors/managers had greater perceptions on all of the dimensions.In other words, the supervisors or managers who are the promoters of establishing a patient safety culture may have increased the outcome of this survey.According to some studies, leaders are the ones who understand current safety issues thoroughly and are able to create a patient safety culture. 21,22hey have a better comprehension of the patient safety culture than others.Their recognition of the patient safety culture will be changed whether or not the hospital policies are practiced well.The dimensions in SAQ-C are jointly related.The overall performance of the patient safety culture is affected by such recognition.Furthermore, some researchers do validate that leadership is associated with safety outcomes. 23A leader can have an immense effect on individual employee behaviors in relation to safety. 24Hospital management should make appropriate interventions to enhance a leader's effectiveness in engaging their subordinates' commitment.
The demographic variable "experience in position" demonstrated a negative influence on the teamwork climate, safety climate, job satisfaction, and perceptions of management.Therefore, employees with less experience in their positions had a higher satisfaction with these four dimensions than those with ample experience.These results contrast with those of Khater et al who explored nurses' perceptions of the patient safety culture. 25Their study indicated that the higher the number of total years of experience is, the better the nurse's perception of the patent safety culture will be.Comparing the current study with the research conducted by Khater et al 25 highlighted some reasons that may have led to the diverse results.First, different measurement tools were used.Instead of the SAQ, Khater et al employed a hospital survey on patient safety culture developed by the Agency for Healthcare Research and Quality (AHRQ).Second, there is disparity in the sample size.Third, the current study surveyed physicians and nurses, while nurses were the only respondents in Khater and colleagues' study.
In the current study, "experience in organization" was shown to have a negative impact on working conditions.Respondents who had less experience in the organization had more satisfaction with the patient safety culture.A research regarding nurses from Korea also showed that nurses with fewer years of work experience at the present hospital held more positive views about the safety culture. 19ifferent age groups revealed diverse perceptions of the patient safety culture.The older healthcare workers in this hospital tended to have a greater patient safety culture.This result was also supported by a research in South Australia. 26It can be inferred that older respondents may have a greater possibility of working in a better working environment, which will increase their perceptions of the patient safety culture.Another explanation might be an old ethic from Asian culture.In the Chinese traditional culture, maintaining harmony is always the top priority in the workplace; this tradition may cause staff members to conceal their true thoughts.Apparently, this might not be happening with the younger generation.
It should be noted that job position only influenced stress recognition.That is, physicians had a better recognition of stress than nurses in this teaching hospital.Considering some aspects from the dimension "stress recognition" on the SAQ, it is interesting to note that this dimension differs slightly from the others.The other dimensions inquire about behaviors that are related to other people in the workplace, while stress recognition focuses on respondents themselves, e.g., personal behaviors.When respondents are inclined to approve the items of stress recognition, that demonstrates how stress affects their work performance.Without a doubt, negative emotions will increase the possibility of workplace deviations.Some organizations or companies do not allow personal emotions to be brought to work, especially negative ones.The regulations or organizational culture mentioned above will affect the replies of respondents.
One possible solution to relieving negative emotions for physicians and nurses is to provide for them the mindfulness-based stress reduction program created by Dr. Jon Kabat-Zinn in 1979, building up the "positivity currency" of the staff, or establishing "appreciative inquiry" groups. 18The general purpose of this program is to teach staff how to deal with stress, pain, and illness.In addition, when individuals focus on positive attitudes toward interactions, life events, and memories, and express their gratitude, "positivity currency" can be printed and stored as assets.That is, maintaining a positive outlook and regularly expressing gratitude are the real values needed for employees to build resilience.Finally, "appreciative inquiry" is a model to cultivate employees in self-determined changes.Praising staff members for their past successful experiences is positive encouragement for employees to enthusiastically enhance resilience and a positive psychology toward negative emotions.

Conclusion
Many countries are making major efforts to enhance patient safety, and healthcare organizations have been encouraged to assess the existing culture in recent years.Such an assessment would provide significant values for enhancing the perceptions of the patient safety culture.It would give hospital management the clues they need to identify flaws in their hospitals and redesign systems if necessary.The current study revealed that "supervisor/ manager" and "experience in position" followed by "age" are the most important demographic variables influencing the patient safety culture from the viewpoints of physicians and nurses.Hospital management should heed the advice of staff members regarding these characteristics.Programs can be established for pioneering patient safety and keeping track of the outcomes from a survey based on the SAQ-C once a year to continuously improve the patient safety culture for healthcare organizations.Specific activities such as a mindfulness-based stress reduction program, positivity currency, or appreciative inquiry groups can be applied to teach physicians and nurses how to strengthen resilience and positive psychology toward negative emotions.Finally, the patient safety culture as viewed by the staff should be tracked on a yearly basis to monitor performance trends in order for healthcare organizations to provide a better healthcare environment for their patients.

Table 1 .
Chinese Version of the 2014 Safety Attitudes Questionnaire

Table 2 .
Demographic Information of the Studied Regional Teaching Hospital

Table 2 .
Women represented 84.8% of all the respondents.A total of 35.9% of respondents were within the age range of 31 to 40 years, while 38.3% of respondents were 21 to 30 years old.Most of them were not supervisors

Table 3 .
Linear Regression With Forward Selection of 7 Dimensions

Table 4 .
Summary of 7 Dimensions and Critical Demographic Variables