Comprehensive interventions of safety culture to improve patient safety: A longitudinal follow-up in northeast China

Background: Safety culture in hospitals can affect patient disease processes and health status. However, comprehensive measures to improve patient safety and effective methods to precisely assess the outcomes are limited in China. Methods: A cohort study was carried out in a tertiary hospital in China. Medical caregivers received comprehensive interventions such as a study on emergency plans and professional skills training to improve safety culture at the beginning of January 2017. A total of 553 caregivers in March, 299 caregivers in June and 284 caregivers in December in the same year participated in the three evaluation surveys. The safety attitude questionnaire (SAQ), the dimension frequency and number of events reported in the hospital survey of patient safety culture (HSOPSC) and a questionnaire on barriers to reporting adverse events were used to examine the safety culture changes before, during and at the end of intervention implementation. Results: The scores for dimension teamwork climate, job satisfaction and perception of management in the SAQ were significantly increased (p < 0.05). The scores for the 17 items in the barriers to reporting adverse events questionnaire were significantly increased (p < 0.05). No significant changes were observed in the scores for the dimension frequency and number of events reported in the HSOPSC after the interventions. Conclusion: Our findings showed that the interventions improved safety attitudes, and the barriers to reporting adverse events decreased, suggesting that the comprehensive interventions used were helpful for improving the safety culture.

in December in the same year participated in the three evaluation surveys. The safety attitude questionnaire (SAQ), the dimension frequency and number of events reported in the hospital survey of patient safety culture (HSOPSC) and a questionnaire on barriers to reporting adverse events were used to examine the safety culture changes before, during and at the end of intervention implementation. Results: The scores for dimension teamwork climate, job satisfaction and perception of management in the SAQ were significantly increased (p < 0.05). The scores for the 17 items in the barriers to reporting adverse events questionnaire were significantly increased (p < 0.05). No significant changes were observed in the scores for the dimension frequency and number of events reported in the HSOPSC after the interventions. Conclusion: Our findings showed that the interventions improved safety attitudes, and the barriers to reporting adverse events decreased, suggesting that the comprehensive interventions used were helpful for improving the safety culture. Background 3 Patient safety refers to a risk reduction in unnecessary harm associated with healthcare to an acceptable minimum [1] . As many as 251,000 people die in hospitals due to medical errors annually in the United States [2] . This results in a total cost of between 17 billion and 29 billion dollars per year. However, some of these costs could have been prevented or reduced to a low level. Another study from Britain in 2000 showed that the incidence of medical adverse events in hospitalized patients was up to 10%, with an economic loss of 3.2 billion pounds [3] .
In China, at least 2.5 million people of 50 million hospitalized patients have adverse drug reactions annually, and medication errors account for 12% to 32% of all drug users. Therefore, the additional suffering and medical costs are significant in these patients [4] . Thus, special attention should be paid to patient safety.
Currently, diverse interventions have been implemented to improve patient safety, and many questionnaires have been constructed to evaluate their efficiency. Paine et al. used a unit-based safety program (CUSP) throughout an academic medical center and showed that the scores of the dimensions in the safety attitude questionnaire (SAQ) improved, except the stress recognition dimension [5] . Haynes et al. implemented the checklist-based surgical safety intervention in operating rooms, and the results showed that the SAQ scores significantly increased, followed by a significant decrease in postoperative morbidity and mortality [6] . Only one study implementing CUSP in twelve departments of a tertiary hospital for 1 year in China indicated that the percentage of positive responses to dimensions in the SAQ increased, except the stress recognition dimension [7] . Thus, a follow-up study of the implementation of comprehensive interventions in a tertiary hospital was carried out to improve the patient safety culture and decrease adverse events. In the present study, we used three quantitative questionnaires to evaluate the effect of the comprehensive interventions over a 10-month period to provide evidencebased data and practical experience to improve patient safety.

Design
This study describes a series of interventions related to patient safety implemented in Shengjing Hospital in 2017. Comprehensive interventions of safety culture were conducted from the first day of 2017. A total of 1066 subjects from several departments were enrolled. Quantitative questionnaires were used to interview the caregivers in the same departments on the last day of March, June and December in 2017.

Setting
Shengjing Hospital is the affiliated hospital of China Medical University (http://www.sj-hospital.org/), and is the biggest comprehensive, tertiary and teaching hospital in three provinces (Heilongjiang, Jilin, Liaoning) in northeast China. In this study, the executive department established the implementation plans, set up the leadership committee and arranged the activities. Moreover, related deans in each department were the persons in charge of the patient safety interventions. To build the idea of a safety culture, the propaganda department used magazines, TV, publicity boards and news media as propaganda tools. The themes of the plan included medical safety, nursing safety, student safety, teachers' safety awareness, research safety, and monthly logistic service safety. Therefore, all these activities ensured the leadership and execution in the study.

Participants and Inclusion Criteria
All staff members were encouraged to study the contents of the interventions and put them into practice. Participants who had worked in their departments for at least one year before the intervention were included in the study, while those who had worked for less than 1 year or left during the interventions were excluded.

Interventions
The scope of these interventions covered the entire hospital. The intervention items in several managerial, clinical and technical departments were introduced ( Table   1). Briefly, we carried out safety culture education, safety knowledge and operation practice with each department as a unit.

Evaluation of the interventions
In this study, three types of questionnaires were introduced [Safety Attitudes Questionnaire, frequency of events reported and number of events reported in Hospital Survey of Patient Safety Culture (HSOPSC), and barriers to incident reporting questionnaire]. A simple random sampling method was used, and the evaluation process was divided into three parts, the beginning, the middle and the end of the intervention. At each time point, we investigated the caregivers who worked in the selected departments. The caregivers who participated in each survey met the following criteria: he or she had worked in the selected department for at least 1 year before the interventions, and persons who left during or at the end of the interventions were excluded. The three questionnaires introduced were as follows: (1) The SAQ which was described previously [11,12] . The SAQ developed by Sexton JB had six dimensions and 32 items in total, and the dimensions were teamwork climate, safety climate, job satisfaction, stress recognition, perception of management and working conditions. All the items were scored on a five-point Likert scale (1=disagree strongly, 2=disagree, 3=neither agree nor disagree, 4=agree, 5=agree strongly), and negatively worded items were reversely scored with higher scores indicating better safety attitudes. (2) The HSOPSC, which was developed by American Healthcare Research and Quality. The Frequency of Events Reported section had three items and they were scored using a five-point Likert scale (1=never, 2=rarely, 3=sometimes, 4=often, 5=always). The Number of Events Reported section had one item with six options: 1=no events reported, 2=1 to 5 events reported, 3=6 to 10 events reported, 4=11 to 15 events reported, 5=16 to 20 events reported, and 6=21 or more events reported. Higher scores indicated a high frequency of events reported and better patient safety. This questionnaire is described elsewhere [13] . (3) Barriers to incident reporting questionnaire. This questionnaire was developed by Vincent and adapted by Evans [14] . The adapted version has nineteen items, whereas in this study we used the version translated by Tian Huanhuan and added the last three items from Evans. All 22 items were scored on a five-point Likert scale (1=disagree strongly, 2=disagree, 3=neutral, 4=agree, 5=agree strongly), in which item 15, 16, and 19-22 were negatively worded, and were reversely scored. Higher scores indicated fewer incident reporting barriers and better patient safety.

Statistical analysis
In this survey, we invited investigators to the selected departments. They addressed the survey and provided the questionnaires in the morning meeting, respondents were asked to complete the questionnaires, and the investigators then collected the questionnaires. In some cases, the head nurses in each selected department provided and collected the questionnaires completed by the few respondents who did not attend the meeting. The interview period for each survey lasted 2 weeks, and a telephone reminder was sent to non-respondents after one week.
The chi-square test was used to compare differences in the composition ratio (%) of demographic characteristics between different months. The Komogorov-Smirnov test was used to identify whether the data were normally distributed. If the data were not normally distributed, the median and interquartile range were used to describe the data. The Kruskal-Wallis H test was applied to compare the scores of the scales with non-normal distribution. The Mann-Whitney U test was used to compare the scores of the scales from two independent samples with non-normal distribution. A p-value less than 0.05 was considered statistically significant.

Demographic characteristics
As shown in Table 2, there was no overall significant difference in the composition ratios of respondent demographic characteristics between the three time points during follow-up, such as age, sex, position classification and job title. Furthermore, no significance differences were found between each two time points during followup.

SAQ assessment
We compared the scale scores obtained in March with those in June (Table 3)..The scores for teamwork climate, safety climate, job satisfaction, perception of management and the total scale dimensions in June were significantly higher than those in March. However, we found that the scores for teamwork climate, job satisfaction and perception of management dimensions in June were significantly higher than those in December. Furthermore, we compared the scores between March and December.

Frequency of adverse event reports and barriers to incident reporting questionnaire
The results in Table 4 showed that there was no significant difference in the frequency of adverse events reported in December compared with that in March and June. The results in Table 5 show that 15 of 22 items had significantly higher scores in December as compared with those in June and March.

Discussion
In the present study, an almost 9-month follow-up period was carried out at Shengjing Hospital from March to December in 2017, and it was found that the safety culture after the interventions had improved markedly using multiple questionnaire assessment.
The scores for teamwork climate, safety climate, job satisfaction, perception of management and total scale dimensions in June were significantly higher than those in March, indicating that the safety interventions had a remarkable effect. However, the scores for teamwork climate, job satisfaction and perception of management dimensions in December were significantly reduced compared with those in June, which indicated that the intensity of the interventions needs to be increased and 9 the intervention time needed to be extended to achieve satisfactory results. The reason for this may be due to seasonal changes, the number of patients and the various disease stages, and these factors may have affected these dimension scores. Scores for the five dimensions, except perception of management and the total scale in December, were higher than those in March, demonstrating that the interventions enhanced the safety culture. The results for teamwork and safety climate were similar to those in previous studies. The comprehensive interventions in this study were similar to those in Paine's study [5] , and our results demonstrated the success of the interventions.
The study by Vigorito [15] included changes in the central line-related blood infection rate and ventilator-associated pneumonia incidence before and after the intervention, but in our study no other patient safety indices were observed. Hefner et al. [16] implemented an intervention on crew resource management, and Gupta et al. [17] implemented the TeamSTEPPS intervention; however, these two methods were not implemented in the present study. With regard to stress recognition, Meurling [18] implemented the systematic simulation-based team training intervention, which significantly increased the stress recognition scores, and no significant differences were observed between any two time periods, suggesting that managers need to implement special interventions for caregiver stress in order to reduce stress and improve the safety culture. With respect to management of the hospital, the score in December was significantly lower than that in June and March.
This indicated that management adjustments were required in accordance with changes in the actual situation in a timely manner, and managers needed to strengthen the intensity of management interventions to improve safety culture.
With the exception of perception of management, the results of this study were consistent with the study by McGuide et al. Their study implemented an electronic medical records system intervention the same as the system in this study, and both studies demonstrated the success of the intervention [19] . Furthermore, the interventions in this study included the use of a surgical checklist, and it was found that after the interventions the SAQ scores increased which was similar to that in the study by Haynes [6] . The interventions in this study included a three level round system and difficult case discussion system, which were similar to the structured interdisciplinary rounds in Leary's study. Both studies showed a higher teamwork score after the interventions, suggesting that the interventions were effective [20] .
In addition, the results of this study indicated a decrease in incident reporting barriers and a satisfactory effect of the interventions. The scores for item 5, 15, 16 and 19 decreased after the intervention, indicating that incident reporting training should be strengthened, and the principal in charge of reporting incidents should understand the concepts involved, such as knowing caregivers' needs, providing the most convenient reporting methods and training on the scope and content of the incidents to be reported. The scores for item 15, 16 and 19 decreased as they were negatively worded, suggesting that these three items should be positively worded to facilitate an accurate answer by the respondents. The findings on safety attitudes and incident reporting in this study were similar to those in other studies.
Caregivers are concerned that the reputation of individuals and hospitals would worsen, which may lead to punishments by organizations as they may be involved in unnecessary lawsuits or disputes if patients and their families are aware of the facts. On the other hand, physicians are reluctant to report adverse events including the risk of liability exposure, professional embarrassment, burdensome reporting methods, time required for reporting, the importance of adverse events and a lack of sense of ownership in the process [21] . Many factors can also influence the reporting of adverse events, such as a shortage of encouragement by the faculty and prompt high quality feedback on adverse events, fear of compromising one's career or personal reputation, time constrains, complex reporting systems, and forgetfulness [22] . The first step should be to create a non-punitive safety culture, improve the mechanism of medical dispute mediation and medical risk sharing. Hospitals also need to strengthen training on adverse event reporting, so that caregivers can report incidents without fear, and safety culture can be improved.

Conclusions
The safety culture interventions in Shengjing Hospital were successful, as teamwork climate, safety climate, job satisfaction and working conditions improved. However, the intensity of the interventions should be strengthened and the time should be prolonged to further improve safety culture. The stress recognition score showed no obvious change, and the perception of management score decreased. These findings indicated that management should be adjusted and strengthened in future. The number of adverse event reports increased and reporting barriers decreased, suggesting that a non-punitive safety culture should be created, the mechanism of medical dispute mediation and medical risk sharing should be enhanced, adverse event reporting training should be strengthened, and caregivers should be encouraged to report incidents, which would improve safety culture. 12 Abbreviations SAQ:safety attitude questionnaire; HSOPSC:hospital survey of patient safety culture; CUSP:used a unit-based safety program.

Ethics approval and consent to participate
This study was approved by the institutional review board of the Shengjing Hospital of China Medical University (2018PS21K). All participants in the study provided written informed consent.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
There are no conflicts of interest. [22] Jansma JD, Wagner C, ten Kate RW, Bijnen AB. Effects on incident reporting after educating residents in patient safety: a controlled study. BMC Health Serv Res 2011, 11:335. Tables   Table 1 Interventions implemented in some departments to improve patient safety implementing "Ten Patient Safety Targets" [8] . key of security checks. Used the information pla anonymously, the e-mail address is shengjingh     When a mistake is made, but is caught and corrected before affecting the patient, how often is this reported?
2 When a mistake is made, but has no potential to harm the patient, how often is this reported?
3 When a mistake is made that could harm the patient, but does not, how often is this reported? 4 In the past 12 months, how many event reports have you filled out and submitted?