Patients and relatives as auditors of safe practices

Background When there is a gap in professionals’ adherence to safe practices during cancer treatment, the consequences can be serious. Identifying these gaps in order to enable improvements in patient safety can be a challenge. This study aimed to assess if cancer patients and their relatives can be given the skills to audit four safe practices reliably, and to explore whether they are willing to play this new role. Methods We recruited 136 The study comprised two parts: an interventional educational program and a cross-sectional design to collect data on the participants’ skills and perceptions about their willingness to be auditors.

educational level, previous experience of an adverse event and perception of hospital safety as predictor variables. Additionally, 106 (77.9%) participants said that they would like to anonymously audit the professionals' compliance of at least three of four safe practices. The willingness to audit safe practices was differed depending on the safe practice but these differences did not reach statistical signi cance.

Conclusion
The data gathered by patients and relatives acting as auditors can provide healthcare organizations with valuable information about safety and quality of care that is not accessible otherwise. This new role provides an innovative way to engage patients and their families (P&Fs) in healthcare safety where other methods have not had success. The paper sets out the methods that healthcare organizations need to undertake to enrol and train patients and relatives in an auditor role.

Background
Improving patient safety is one of the major targets for healthcare organizations. Studies have shown that adverse events in hospitals can vary between 3-17% of all hospital admissions. 1,2,3 . World Health Organization (WHO) safety programme and Spanish Ministry of Health safety programme; as well as other European organizations like the European Patient Safety Foundation (EPSF), medical societies and healthcare providers, encourage governments to develop tools to ensure that healthcare is both safe and patient-centered. 4 There has been quite a lot of work to engage patients in safety, for example in medication safety. The WHO Alliance for Patient Safety also emphasized that the patients' family members could play an important role in the improvement of care. Nevertheless, despite the emphasis and importance of patients' involvement in promoting safety and reducing adverse events, there has been insu cient progress in this area. 5,6 .
Cancer care is complex and requires particular effort to assure safety in care delivery for patients. It is known that cancer patients are vulnerable to breaches of safe practices because of their health conditions (for example immune suppression) and the nature of their treatment. Retrospective chart audit studies of acute care in several countries have shown that between 3% and 16% of patients experience one or more harmful adverse events while hospitalized and that about half of these events are preventable 7 . The areas of cancer care where gaps in adherence to safe practices can lead to adverse events with potentially serious consequences include patient identi cation; correct choice, dose and route for the delivery of chemotherapy and transfusion medication; and infection control. It is well established that patients and their families (including friends and informal caregivers) have unique knowledge and are able to detect if their care is safe and patient-centred. 6,8,9 Patients and their families are present during the whole care episode and often are the only members who are aware of lapses in safety thus being a useful source of information about patient safety.
Information gathered by patients and their families (P&Fs) gives healthcare organizations an opportunity to learn and improve the system of care. 10,11,12 One way of playing the auditor's role is through the patient-as-observer approach. 13 This approach involves recruiting a cohort of patients with multiple healthcare contacts who report on a continuous basis whether health professionals correctly follow patient safety protocols. However, there is little evidence collected directly from patients about their willingness or ability to be involved in this new patient safety role. 14,15 The aims of this study were to assess if in controlled conditions, P&Fs can a) reliably audit safe practices; b) be given the skills and thus describe the characteristics of good auditors; and c) to explore if P&F's are willing to play this new role. Based on the evidence from real-time safety audits performed during routine work, it is known that such audits can detect a broad range of errors. 12 From this it was considered that the patients' participation in the role of an auditor could assist in identifying gaps in safety and this could lead to work to improve patient safety 6,12 . The safe practices selected for evaluation were patient identi cation, hand hygiene, blood or chemotherapy identi cation, and secondary effects of chemotherapy/transfusion. These practices were selected as they can be observed by P&Fs and because of the serious risks to patients if the protocols are not followed.

Materials And Methods
The methods involved an interventional educational program to improve the participants' skills to audit safe practices and a cross-sectional study using a questionnaire to collect data on the participants' perceptions about their willingness to be auditors. The opinions of the patients provided in two focus groups were considered in designing and executing the study. These focus groups were organized during the development phase of the study and took place during the rst months after funding.

Participants
A consecutive sample of 136 participants was recruited between March and October 2018 from the oncology and haematology day hospital of a tertiary hospital in Spain. Patients were eligible to participate if: they were older than 18 years; it was not their rst treatment appointment; the treatment lasted several hours; the healthcare professionals in charge considered their physical and psychological status as acceptable for participation; and they were able and willing to give their informed consent to participate.
Patients' family members were recruited after being informed about the study and providing their consent to participate.
For this study, we considered as relatives not only family but also friends or informal caregivers. The research was conducted while patients were being treated in order to emulate real conditions. We selected day hospital oncology P&Fs because they have multiple contacts with the healthcare organization and thus it can be feasible to train them. Likewise, non-compliance of professionals with safe practices can have serious consequences for the patients' health. The ethics committee's approval was obtained, and all participants provided written informed consent.

Methods
The videos and training lea ets were produced at hospital by personal working on the study or volunteers that collaborate as actors in our medical installations. Also, the digitization unit help us with videos recording and edition.
Initially, the P&Fs watched an assessment video only once. Immediately after viewing it, they lled an evaluation grid.
Then, they were provided with a training brochure. Afterwards, participants watched a training video. Participants could watch this last video as many times as they wished. After reading the material and watching the videos, the participants once again watched the assessment video and lled a second evaluation grid. Later, they completed the questionnaire about their willingness to become auditors. The participants could ask questions during this process. Depending on the time availability of the participants, debrie ng was done once the process had nished. All the videos were played on a tablet. The whole process lasted between 60 and 90 minutes.

Materials and procedure
The research materials comprised: training brochures, videos (assessment and evaluation), evaluation grids, and a questionnaire to assess P&Fs' willingness to audit. The training brochures explained how healthcare professionals must implement the four safe practices selected for evaluation. The videos were lmed in the real places where patients receive treatment. There were two different stories, depending on the type of P&F. They showed a patient who goes to the day hospital to receive chemotherapy or to receive transfusion. Both videos were in a story-like format intending to show, in the most realistic way, the interaction between a patient and healthcare professionals during treatment. The right way to implement the safe practices was highlighted in the training video. The actors were different in all videos in order to make a distinction between the training and the assessment video. Additionally, the content of the materials used were adapted according to the treatment (oncology or haematology).
The evaluation grid had 7 or 8 questions depending on the treatment (Fig. 1). The questions dealt with the observation of the ful lment of the four safe practices studied in the video.
All the materials were pretested with 48 convenience individuals to ensure that all the materials were easily comprehensible. Their answers helped to test the material but were not included in the results. One member of the research team, after explaining the purpose of the study, guided each participant through every step of the study.

Measures
A variable named "potential auditor" was created in order to analyse the P&Fs' degree of willingness to audit healthcare professionals. Participants who answered that they would audit at least 3 out of 4 safe practices were considered potential auditors. A variable called "good auditor" was also created to measure participants who correctly answered more than 75% of the items. This cut-off point was chosen based on the margin of error that the organization was willing to assume.
The dependent variables were: "potential auditor" and "good auditor." The independent variables were gender, age, type of participants, type of treatment, number of healthcare encounters, adverse events suffered, education level, and general perception of hospital safety.

Data analyses
Data were analysed using SPSS version 20.0 for Windows. Wilcoxon and McNemar tests were used to compare beforeafter results; Pearson's chi-square test or Fisher's exact test was used for discrete variables and Student's t-test for continuous ones; and logistic regression (forward stepwise) was used to measure the in uence of different variables on P&Fs' willingness to audit and being a good auditor.

Participant characteristics
In total, 136 P&Fs agreed to participate (63% response rate). The characteristics of the participants included age, gender, education level, number of health encounters during last year, type of participants, type of treatment, healthcare professionals, if they have suffered any adverse event, and perception about hospital's safety ( Table 1).
The proportion of patients (66%) was much higher than that of relatives because often patients come to the day hospital on their own. Almost half of the P&Fs had a basic level education (just primary and secondary school studies) and the others university or superior studies. The proportion of women and men was very similar. As expected, most P&Fs were not healthcare professionals although 14 were. The number of healthcare encounters (mean: 10 hospital visits and 1 hospital stay) indicated that the P&Fs had plenty of experience of visiting the hospital.
Ability to recognize safe practices after training Of the P&Fs, 88.6% answered that they felt con dent in identifying 3 out of 4 practices after the training.
Participants' skills to be auditors The variable good auditor was measured before and after the training to analyse the baseline skills of the participants to be auditors and the improvement achieved after the training. The overall percentage of this variable increased from 30.4% before training to 46.3% after it (McNemar p=0.000). P&Fs had similar baseline skills (30% patients vs. 30.4% relatives). After the training, relatives improved more than did the patients (58.7 % vs. 40%; p bilateral exact Fisher= 0.046)

Characteristics of good auditors
Sixty-three participants (46.3%) were classi ed as good auditors after the training (% of correct answers > 75%). Age, education level, type of participants, type of treatment, adverse events, and general perception about hospital's safety showed statistically signi cant differences in the bivariate analysis tests ( Table 2).
A multivariable logistic regression model was developed to simultaneously consider all the variables in order to predict the outcome of "good auditor." The variables age, adverse events, hospital safety perception, and education level had statistically signi cant coe cients (see Table 3). Younger age, high educational level, experience of an adverse event and not having an excellent general perception about hospital's safety are the characteristics that best predict a P&Fs being a good auditor. The value that indicates the area under curve was 0.838 indicating that the variables included in the model have a strong effect (see Figure 2).

Willingness to be an auditor
The percentage of potential auditors varied for each safe practice. It was 72.1% for hand hygiene, 75% for secondary effects of chemotherapy/transfusion, 79.1% for patient identi cation, and 80.1% for blood or chemotherapy identi cation. Moreover, 106 participants (77.9%) said that they would like to anonymously audit the professionals' degree of compliance for at least 3 out of 4 safe practices. Only the number of hospital visits was statistically different among participants who were willing to audit and those who were not in the bivariate analysis tests (Table 4).
In the logistic regression analysis, none of the variables included had statistically signi cant coe cients.
Among 104 participants who were willing to participate, 43.3% were good auditors. Of 63 participants categorized as good auditors, 45 (71.4%) were willing to audit.

Discussion
It is known from research that there is a gap in professionals' adherence to safe practices 6 . Some authors highlight that a video could be an important educational tool for increasing patients' knowledge of the role they can play during hospitalization 16 .
Walsh et al. estimated the rate of error in the administration of chemotherapy as 8.2 per 1000 orders in oncology adult patients in the outpatient setting, causing damage in one error for every 1000 orders. 17 A considerable fraction of these occur in the phase of administration, which is observable by patients. Qualitative studies have shown that patients are aware of the medication errors that are occurring and are prepared to participate actively in their prevention. 18,19 Also, some works suggests that healthcare professionals, like patients, generally view patient involvement positively. 20 The National Haemovigilance Report published in Spain, in 2016, recorded 332 errors in the administration of blood products, and 32% of them happened at "the bedside of the patient." 21 However, studies on patients' involvement in transfusion safety are scarce. 22 Between 5% and 10%, patients admitted to a hospital will develop at least one nosocomial infection. Hand hygiene is one of the main measures to prevent these infections 2 . However, according to international studies, the adherence of professionals to hand hygiene is less than 50%. 15 Multimodal strategies are being implemented to improve their adherence. These strategies have had a variable effectiveness (51-83%) and there is some evidence that patients can play an important role in improving the compliance. 23 Proper patient identi cation at every step of clinical care is vital to ensure patient safety. However, despite the priority placed on addressing this issue, signi cant problems persist. "Wrong patient," "wrong site," and "wrong procedures" continue to be among the most frequently submitted sentinel events reported to The Joint Commission in the USA. 24 These data suggest that new methods are needed to assess professionals' adherence to safe practices. Such data has the potential to allow clinical teams and services to consider the reasons for non-adherence and to make changes to improve patient safety.
Our research showed that more than 3 out of 4 participants were willing to play the role of the safety auditor. Several reasons can explain this nding, some of which are related to the process of reporting data to healthcare organizations.
In the rst place, we proposed that the assessment must be done anonymously so the participants would not have to confront healthcare professionals. Second, the participants received a training inducing more con dence in their skills.
In fact, 89% participants answered that they knew how to assess safe practices after the training offered. Other reasons are related to the participants' characteristics. Oncology patients may perceive a high risk of an error and thus be more willing to play an active role in patient safety. All these are enablers of patient involvement in patient safety 14,25−29 . In the study P&Fs' participation was intended to simulate a continuous assessment during their process of care. Their willingness here may be different to a more conventional audit team, in which the P&F is a member along with healthcare professionals.
On the other hand, our research showed that the willingness to audit safe practices was different depending on the safe practice and whilst these differences did not reach statistical signi cance, it is interesting to note which practices were selected. Transfusion or chemotherapy identi cation were the safe practices that P&Fs were more willing to audit while hand hygiene was the least selected practice. Somehow there many reasons that could in uence P&Fs preference and willingness to engage in their healthcare. Some studies revealed that there is a general expectation that healthcare professionals, "know what they are supposed to be doing" and a common assumption that they always did what they were supposed to do, specially the most basic duties as washing their hands properly or administering the correct medications. Also, some studies suggest that checking to ensure that healthcare professionals were doing their job correctly could be embarrassing and damage relationship with them.
Participants were offered training before assessing their observation skills to increasing their health literacy. After the training, almost half of participants were considered to have the skills to be an auditor. It means that not every P&F willing to be auditor could or should be. Younger participants with high education level who have experienced an adverse event or who did not rate "totally safe" in the item General perception of hospital's safety made the best auditors. Our study showed that P&Fs with higher education status, a proxy for health literacy, 30 were better equipped to identify noncompliance with safe protocols. In this study, relatives had better skills to play the role of the auditor in the bivariate analysis and they improved after the training more than patients did but the result did not reach statistically signi cant differences in the regression model, probably due to the relatives' sample size.
Here we recommend that organizations develop methods to assess the skills of P&Fs before they are fully engaged in this audit process. Further research is also needed to assess and develop training programmes for patients and family members as auditors.
P&Fs' assessment of gaps in safe practices gives the organization real-time data in order to engage them in the plan-docheck-act cycle. Furthermore, the fact that professionals may feel observed could encourage their adherence to safe practices.

LIMITATIONS
Although P&Fs as well as the environmental frame are real-world entities, the evaluation of the professionals' safe The evaluation of the role of the P&Fs as auditors of safe practices is an innovative approach. Therefore, from an ethical point of view, it seems more reasonable to assess safe practice under real conditions only if minimum guarantees of success are met. Further, the acceptance of the role of P&Fs as auditors implies, not only that the P&Fs themselves have accepted this function, but also that the healthcare professionals and the management team also accept it. Using data to demonstrate that P&Fs are able to audit correctly facilitates the acceptance of this new role by health professionals and P&Fs.

Conclusions
Using P&Fs as auditors of safe practices has many advantages. It goes beyond P&Fs giving an opinion or lling a perception questionnaire. Auditing goes directly to the quality assessment of healthcare organizations. 28 Using patients and family members in auditing allows for continuous monitoring and highlighting the importance of the inherent and essential agents in every healthcare process: the patient and the companion. Furthermore, it allows organizations to assess areas and departments that it would be otherwise impossible or that would require extraordinary efforts.
This new role has advantages not only for the organization but also for the P&Fs themselves. In order to play this new role P&Fs have to acquire the necessary skills. These new skills can enable them to adopt more active behaviours towards professionals such as "speaking up" and thus add new safety layers in providing a safer care. 15 Three out of four participants recruited were willing to play the safety auditor role. Organizations that want to implement this strategy have to carefully select candidates.
Patients' participation in auditing safe care can be an innovative and viable approach to helping organisations improve the safety of the care they deliver.   Figure 1 Items included in the evaluation grid.