Improving uptake of guidelines

Translating guidelines into clinical practice is challenging, and simple guideline dissemination is unlikely to change clinical practice behaviours that are influenced by a multitude of factors [20]. It is important to understand what is driving behaviour when developing interventions to change it [21,22]. Ideally, this process should be guided by a theoretical framework of established psychological theories of behaviour change [20–23]. The Theoretical Domains Framework (TDF) was developed by Michie et al. to identify factors influencing health professionals’ implementation of evidence-based guidelines into practice [24–26]. The TDF simplifies 128 constructs from 33 behaviour change theories into 12 theoretical domains [26,27]. Each theoretical domain may be a determinant of the behaviour that requires change.

We are aware of only one Canadian study using a theory-based approach to examine healthcare providers’ (HCPs) barriers to de-implementing routine preoperative tests for low-risk surgeries. [28] Patey et al found the most common barriers identified by anesthesiologists and surgeons practicing in Ontario, Canada were related to the lack of clarity regarding who was responsible for ordering the tests, perceived inability to cancel tests ordered by other HCPs, and tests being ordered and completed prior to the anaesthesiologist’s examination of the patient [28]. While some of the drivers of behavior may be similar among HCPs across Canadian provinces, healthcare systems are provincially managed leading to contextual differences in each province. Indeed, even within provinces, practices can vary from hospital to hospital. We completed a TDF-guided investigation of our local context by conducting a series of semi-structured interviews with surgeons, anaesthesiologists, and preoperative clinic nurses in Newfoundland and Labrador (NL) to identify perceived barriers and enablers to reducing unnecessary preoperative tests for low-risk surgical procedures.

Design

This was an exploratory, qualitative study conducted using semi-structured interviews to understand HCPs’ barriers and enablers to de-implementing low-value preoperative testing. The protocol for this study was previously published [29]. This study was reported according to the Consolidated criteria for Reporting Qualitative research (COREQ) checklist [30].

Participants

Participants were selected using a purposive snowball sampling strategy in order to ensure diverse perspectives and representation across the four regional health authorities (RHA) in NL, as well as a variety of surgical subspecialties. We identified a key informant within the largest RHA, Eastern Health, to provide a list of three potential participants as well as potential informants from the other three RHAs across the province. At the end of each interview, participants were asked to identify additional participants who may be interested in participating in this study.

Eligible participants were HCPs (surgeons, anaesthesiologists, or preoperative clinic nurses) practising in NL who order preoperative tests for patients undergoing surgery.

Data collection

Potential participants were emailed by a researcher (KM or AH) to gauge their interest in participating in the study. Oral consent was obtained, and semi-structured interviews were conducted by two interviewers (AP and AH, both females, with graduate degrees in the health sciences, and employed as academic faculty and/or researchers). The interviewers were trained in qualitative methods and interview techniques, with over 15 years of experience. Participants learned about the interviewers at the start of the interview via verbal introduction; there was no relationship between the interviewers and participants established prior to the start of the study. Field notes were taken by a non-participant observer (RL or KM) during the interviews as a fail-safe measure in case of a recording failure) but were not used in the analysis. Interviews were conducted over the phone or in-person. All interviews were audio-recorded and transcribed verbatim. No repeat interviews were carried out. Transcripts were not returned to participants for comments and/or corrections and participant checking of findings was not performed.

Interview guide

The behaviour of interest was ordering of preoperative tests (chest x-rays and ECGs) for healthy patients undergoing low-risk surgical procedures (knee arthroscopy, laparoscopic cholecystectomy, cataract removal, and similar types of surgeries). Healthy patients were defined as those without comorbidities or additional medical conditions that could complicate anaesthesia management and perioperative care [28]. The interview guide (adapted from Patey et al.’s study with surgeons and anaesthesiologists on preoperative testing in Ontario [28]) was developed using the TDF to elicit participant beliefs about their behaviour through the lens of each domain. It included 1–4 questions per domain (12 domains), for a total of 31 questions and prompts were provided in the interview guide to assist the interviewer in clarifying participants’ responses if needed. No changes were made to the guide after pilot testing with our key informant. Please see the published study protocol for a copy of the interview guide [29].

Data analysis

Following the method outlined in the TDF guide [26], coders read and reread transcripts to become familiar with the data. Coding began after six interviews were completed and transcribed. Using the TDF to generate a framework for content analysis, researchers analyzed the data deductively (assigning text to one or more domains) and inductively (identifying themes within each domain.

Identifying relevant domains

Relevant domains were identified through consensus discussion between the two coders (SI, SM), confirmed by AMP, and subsequently reviewed with the larger research team. Relevant domains are those for which sufficient data were coded to inform an intervention for behaviour change [33]. Three factors were considered to identify relevant domains: (1) reported strength of opinion that the beliefs influenced the behaviour, (2) presence of conflicting beliefs, and (3) frequency of the beliefs across interviews [26]. All factors were considered concurrently to establish domain importance.

Ethics

Ethics approval was obtained from the Health Research Ethics Board in Newfoundland and Labrador (HREB #2018.190).

Preoperative assessment process

Participants reported different definitions of “low-risk surgeries,” but generally, both patient characteristics (past medical history and physical exam) and type of the surgery were reported to play a role in defining a surgery as low-risk. Same-day surgeries (i.e., the patients who could go home after the surgeries) and those that did not require general anaesthesia were regarded as low-risk. Some participants considered patients of certain age groups (e.g., 50, 60, 70 years) as high-risk regardless of other factors.

There were inconsistencies among the participants when describing the process for preoperative assessment. Most often, patients were initially seen by surgeons. Based on their medical history, physical exam, and type of required anaesthesia, they were either referred to the pre-admission centres (PAC) or were booked directly for their surgery without any preoperative assessment. If patients were referred to a PAC, they would first undergo an assessment by PAC nurses (sometimes this would be done via telephone). The nurses reviewed patients’ medical history, discussed the details of the surgery with them, and referred them to in-person visits by anaesthesiologists for further examination and tests (if necessary). However, in some centers, the anaesthesiologists’ visits were only required if any red flags were highlighted by the nurses.

Although most participants reported that the tests were ordered at PACs by anaesthesiologists, tests were also ordered at surgeons’ offices or by nurses during the telephone interviews. In some rare occasions, especially if patients were admitted as outpatients by surgeons, the anaesthesiologists would not see the patients until the moment of surgery.

Key themes among relevant domains

Theoretical domains identified as relevant to preoperative test ordering were: beliefs about capabilities, beliefs about consequences, social/professional role and identity, motivation and goals, social influences, environmental context and resources, and behavioural regulation. Table 1 presents a summary of the belief statements and supporting quotes for each of the relevant domains.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of belief statements and sample quotes from anesthesiologist, surgeons and nurses assigned to the theoretical domains identified as relevant.

https://doi.org/10.1371/journal.pone.0278549.t001

Four overarching themes are described below

1. While HCPs may not intend to routinely order preoperative tests, in practice, a host of factors play a role in ordering decisions including use of automatic order sets, workplace norms, expectations about the needs and wants of other HCPs, and lack of access to information, technological tools, and human resources.
   Most participants stated that they did not intend to routinely order tests and were comfortable with the idea of not doing so. However, some believed that the tests are important to carry out to provide the clinical team with as much information about the patient (and potential adverse outcomes) as possible (motivations and goals, beliefs about consequences).
   Participants reported a variety of factors that impacted their ordering decisions, some of which were outside of their influence (e.g., institutional use of automatic order sets and pre-defined templates for preoperative evaluations (environmental context and resources) or tests ordered by other health care providers that they were unable to cancel (beliefs about capabilities). Other factors that influence ordering decisions include perceived needs/requirements of other HCPs, workplace norms, and avoiding inter-personal conflicts (social influences). Fear of having procedures delayed or cancelled by anaesthesiologists if tests were not completed and fear of legal issues if patients were to experience an adverse outcome were also reported to influence ordering decisions (beliefs about consequences). Finally, while participants indicated a lack of environmental constraints to ordering, they did highlight environmental factors that make it more difficult for them to avoid routine ordering including a) a lack of access to technology and/or online tools (which makes it difficult for them to check past medical history of their patients and the results of any tests they have undergone), b) lack of staff and difficulty in accessing other specialists for consults (environmental context and resources).
2. HCPs believe that the positive effects of not ordering outweigh the negative and don’t think the tests are likely to influence their practice but some still consider them a fail-safe–to ensure procedures move ahead as scheduled and that important clinical findings are not missed.
   Most participants agreed the benefits of not ordering tests outweighed any potential risks (beliefs about consequences). Reported positive effects included reduced patient inconvenience, fewer false positives and unnecessary follow-up tests, less paperwork, reduced wait time for surgeries, and higher efficiency and cost-effectiveness (beliefs about consequences). They also stated that the results of the tests were unlikely to change their practice. Nevertheless, some participants noted that failing to order tests may result in surgery delays or cancellations or missing important clinical findings (beliefs about consequences).
3. There is a lack of clarity about who is responsible for deciding whether or not to order preoperative tests, though most agree that chest x-rays and ECGs are not required for low-risk patients.
   There was significant confusion and conflicting comments among the participants with regards to responsibility for preoperative test-ordering (social/professional role and identity). The surgeons we interviewed reported that they have a different set of concerns and goals for surgery (compared to anaesthesiologists) and may therefore want to order different tests to prepare for surgery. They also reported that if they were unsure or concerned about a specific case, they would refer it to their anesthesiologist or internist colleagues for consultation (social influences). While both groups agreed that chest x-rays and ECGs for low-risk patients undergoing surgeries was not an expected part of preoperative assessments, it was not clear who was responsible for deciding whether or not to order these tests. Rather, we found that preoperative test-ordering (meant to support anesthesia management) was not at the sole discretion of the PAC or anaesthesiologists (social professional role and identity). Sometimes surgeons order these tests in an attempt to ward off potential cancellations or delays. In addition, participants reported that tests were sometimes ordered under surgeons’ names by other HCPs without the surgeons’ awareness. Consequently, the HCP ordering the test could not check the test results because they were sent to the surgeons who then became responsible for following up on the tests they had not ordered. A few participants also reported that sometimes surgeons’ administrative staff would order tests on their behalf using order sets for all patients regardless of patients or surgery characteristics.
4. HCPs believe that implementing clearer guidelines, enhancing understanding of current evidence, and building consensus on a preoperative testing strategy, supported by improved communication and access to resources, and periodic evaluation will reduce unnecessary preoperative testing.
   Participants emphasised the importance of clearer guidelines, educating stakeholders on the current evidence, and ensuring that there is consensus among stakeholders regarding preoperative testing strategy to reduce routine preoperative testing (behavioral regulation). They also explained that endorsement of policies and guidelines by regulatory bodies could help reduce inappropriate test ordering. Several strategies were mentioned to help reduce test ordering: performing a thorough clinical history and physical exam, detailed documentation of patients’ symptoms and reasons for testing/not testing, easy access to online tools (e.g., Meditech or HEALTHe NL), enabling pop-up questions when ordering a test (e.g., asking if patients’ symptoms were new), and removing or changing order sets (behavioral regulation). Participants also suggested that audit and feedback on appropriateness of preoperative assessments and reminders about the benefits of not ordering could reduce test ordering. In terms of inter-professional relations, changing the accountability of test ordering and better communication with anaesthesiologists were highlighted.

Themes among domains identified as not relevant

Table 2 presents a summary of the belief statements and supporting quotes for each of the four domains that were not considered relevant to preoperative test ordering: knowledge, skills, memory and decision making, and emotion. Most participants were aware of the preoperative guidelines and believed them to be evidence-based and trustworthy. A small number of participants were not sure about the details of the guidelines (knowledge). Almost all participants believed that basic clinical skills were enough for performing a history/physical exam and deciding whether tests were needed and that experience could improve the skills needed for preoperative assessments. They also felt that good interpersonal and communication skills with patients were important (skills).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of belief statements and sample quotes from anesthesiologist, surgeons and nurses assigned to the theoretical domains identified as irrelevant.

https://doi.org/10.1371/journal.pone.0278549.t002

The decision to not order tests was an easy decision to make by most participants. However, as mentioned previously, some participants recognised that automaticity and environmental factors influenced their test ordering decisions and explained that they were mandated to automatically order tests according to policies set by hospitals, for example, if patients were a certain age (memory and decision processes). Almost all participants mentioned that not ordering tests did not worry them (emotion).

Findings in relation to previous research

We found only three studies investigating why routine testing practices persist despite guidelines that advise otherwise. Two were very similar to our study parameters including interviews with surgeons, anesthesiologists, and/or nurse administrators (conducted in the US and Canada) [28] [34]. The third interviewed general surgeons involved in preoperative clinics in New Zealand [35].

Consistent with these studies, participants in our study agreed with guidelines recommending against low-value, routine preoperative test ordering and did not believe that these tests help or influence their practice [28,34]. Rather, they believe not ordering tests routinely can benefit patients (for example, by reducing patient discomfort and inconvenience) and the health system (for example, by reducing costs and improving efficiency) [28,34]. Additional practitioner-level barriers also noted in this study and the wider literature include the fear of missing “something serious” that could lead to adverse patient outcomes [34,35], medicolegal concerns, [34], worry that procedures will be cancelled or delayed if preoperative tests are not carried out [28,34], and beliefs about the preoperative testing needs and expectations of other specialties [28,34]. Further complicating the issue is confusion about professional responsibility (i.e., lack of clarity on who is responsible for ordering preoperative tests)–an issue also noted by Patey et al [28].

Our findings of systems-level drivers of routine testing were also reflected in the available literature. Brown & Brown (2011) and Raina et al. (2019) also found that practice traditions or workplace norms (e.g., routine testing strategies) were important barriers to reducing preoperative testing [34,35]. Additionally, Raina et al. (2019) also found that poor systems of communication between the HCPs and the departments involved, and a lack of staff or HCP time available for consults [35] are important barriers to change. These barriers, particularly poor communication between professions, is perhaps not surprising or unique to preoperative testing. Communication failures among teams of HCPs have been observed in many areas of healthcare resulting in redundancy, treatment errors, or even patient harms [36–42].

In contrast to other studies in this area [34,35], participants in this study were aware of preoperative testing guidelines and believed them to be evidence-based and trustworthy. This replicates findings from a Canadian study [28] which also didn’t find that knowledge was a barrier to reducing preoperative testing and which had very similar study parameters and used a very similar interview guide. Guidelines recommending against routine preoperative testing strategies have been published for many years [43] and in recent years, Choosing Wisely has been actively campaigning on this issue which may partially explain why knowledge doesn’t appear to be lacking among the participants we interviewed. Also, the questions about knowledge in the US [34] and the NZ [35] studies were focused on familiarity with the evidence around the benefits of preoperative testing, which guidelines were used in decision-making, and, in the NZ study, knowledge of the Choosing Wisely campaign. Globally, then, its not clear that knowledge is a major barrier but it does not appear to be the case in Canada.

Our participants also described additional contextual factors that drive routine testing including automatic order sets or pre-defined templates for preoperative evaluations, as well as lack of access to technology allowing for easy access to patient information. This issue was not initially included in our interview guide but came up in an early interview and was probed in subsequent interviews. This may be an issue relevant only in the local context, or it may not have surfaced previously simply because it was not asked about.

Finally, we found the desire to avoid interpersonal conflicts with colleagues also influenced physicians to order routinely. We believe this finding is related to other common barriers such as HCPs’ beliefs about the needs and expectations of other HCPs and confusion about which HCPs are responsible for ordering. We were aware that previous studies had noted these types of issues so we knew to probe them if they came up in our own interviews resulting in a more nuanced understanding of this issue.

Strengths and limitations

As recommended by numerous health and research organizations, (e.g., the National Institute for Healthcare Excellence, the MRC, Health Canada, Canadian Institutes of Health Research, and the Quality Enhancement Research Initiative) [44–48], this study used a theory-guided investigation of the determinants of routine preoperative testing for low-risk surgeries. Using the Atkins et al guide [26] on how to apply the TDF to guide our assessment and analysis allowed us to produce results that can be used to develop a theory-informed intervention best-suited to tackle known barriers to reducing routine testing. Additionally, we applied rigorous data collection and analysis methods (including a TDF-based question guide, extensive coder training, double coding of all transcripts, oversight of analysis by a professional expert in the TDF, and review of results interpretations with clinical experts and the investigative team. interviews until data saturation, double coding, coding to consensus, interpretation checked by experts and reviewed with team). We also published the study protocol which was reviewed by content experts in the TDF, qualitative research, and preoperative testing and used the Consolidated Criteria for Reporting Qualitative Research (COREQ) 32-item checklist to guide our methods and reporting [30].

Despite this, our results are limited in several ways. We were unable to interview participants working at all sites offering low-risk surgical procedures in the province as originally planned. However, we were able to include participants from all but one regional health authority. In addition, although we reached data saturation, it is possible that we did so prematurely by not interviewing participants with sufficient diversity to allow for more variety in responses. For example, while we were able to speak with HCPs in three of four provincial health regions, we did not interview providers working in Labrador who may face different barriers to reducing routine preoperative testing. Finally, we didn’t assess the test-ordering practices of our sample to engage similar numbers of participants that order these tests at different rates; nor did we actively seek participants in equal numbers who had differing views on routine preoperative test-ordering.

Implications for research and practice

While HCPs interviewed for this study were aware of when to order preoperative tests according to guidelines, there are both system and practitioner-level barriers that continue to influence their practice. Many of these barriers occurred as a result of the complex nature of preoperative care and a lack of communication among providers suggesting that implementation is likely to be challenging. This challenge in implementing guidelines is further complicated when the health system fails to provide a clear set of guiding rules standardized for all professions, lack of communication with and “buy-in” from all professions involved in the patient’s care, and lack of easy access to patient information. Based on these findings in NL, our next steps will be to select strategies for an intervention that can be trialled locally.

Clear endorsement of guidelines by health authorities/hospital administration and development of local policies about who is responsible for ordering tests which are clearly communicated to all surgical HCPs (surgeons, anesthetists, pre-admission clinic staff, and potentially GPs) could address most of the barriers we identified. This includes additional issues stemming from the “just in case” types of ordering which would likely be abated if the direction of the hospital is clear and if HCPs can support their decisions to not order tests using easy to access and up-to-date data from patients’ medical records.

In the literature we found only three studies that have tried to reduce preoperative testing, all of which used strategies similar to what we have proposed above (two in the US [49,50] and one in Canada [51]) with the exception of clarifying responsibility for ordering. All interventions included some form of education for HCPs involved in ordering. In addition, two included decision support tools [49,51] and/or use of champions [50,51] to reinforce educational messages. While all studies showed promising results, certainty of the evidence is low. Trials in this area are long overdue. Recently, Ahmadi et al (2021) published a protocol for a randomized superiority trial planning to implement educational messages triggered by preoperative test ordering the EHR [52]. In Canada, a trial has been registered that aims to reduce preoperative testing for low-risk ambulatory procedures by improving accountability for ordering. They will do this by implementing a hospital policy directing preoperative tests to be ordered only by consulting anesthesiologists based on their clinical assessment of the patient undergoing surgery [53].