We used semi-structured interviews to address the study objectives and gain insight into the factors influencing the use of LVC, the magnitude of the problem of LVC, and unique aspects of de-implementation processes. A qualitative methodology is an ideal approach to explore perceptions and experiences of a phenomenon [24]. The Standards for Reporting Qualitative Research (SRQR) were used to guide the reporting of this study (See Additional file 1). This study has received ethics approval from the Office of Research Ethics at the University of Toronto (Protocol # 00038952) and the Office of Research Ethics at North York General Hospital (Protocol #20 − 0019).
Interventions related to Choosing Wisely Canada recommendations in hospital and primary care settings in four Canadian provinces.
We used a purposive recruitment strategy. The inclusion criteria were: team lead or member that implemented an active intervention to address a Choosing Wisely Canada recommendation in the last five years. Our goal was to interview 15–20 participants based on feasibility and recommendations from the literature that qualitative studies typically reach data saturation after twelve interviews [25]. Participants were recruited through two methods. Initially, participants were recruited by a purposive sampling strategy through their participation in a hospital Choosing Wisely Canada Committee. Potential participants were contacted via email and were sent an email invitation that contained a short description of the participation requirements, expectations of participants, and a brief recruitment survey. One-third of participants (n = 6) were recruited through the initial strategy. Due to the impact of the Covid-19 pandemic on the availability of potential participants, we expanded the recruitment strategy to identify additional potential participants from individuals who participated in the Choosing Wisely Canada Annual National Meetings in 2018 & 2019. These potential participants were invited to participate in the study via their publicly available email addresses. All interested participants responded and provided high-level details about the Choosing Wisely Canada intervention they were involved in and their availability through the recruitment survey. Our goal was to obtain a diverse sample across hospital and primary care, geographic locations and targeted LVC. From the respondents, those who met the inclusion criteria were selected for an interview.
Interviews were selected as the ideal data collection method as they facilitate the exploration of context and the complex interplay between individuals, processes, and structures [26]. The interview guide was developed using concepts from the Framework for de-implementation in cancer care delivery [1] and the Implementation Process Model (IPM) [23]. In their framework, Norton, Chambers & Kramer (2019) detail numerous factors in the Continuum of Factors Influencing De-Implementation Process particularly the conceptualization of the Magnitude of the Problem of LVC. The IPM details key elements in the intervention implementation process, and questions were developed regarding the pre-implementation (planning) phase, implementation phase, monitoring, and evaluation. Interviews concluded with questions about participants’ experiences with de-implementation and how it differs from implementation efforts. In addition, the guide gathered data on pre-intervention rates, targets, intervention strategies, and outcomes. The interview guide was iteratively developed by the research team and finalized at 15 questions (see Additional file 2).
Interviews were approximately one hour in length led by an experienced facilitator (XX) who had no affiliation or existing relationship with eligible participants. Due to the in-person meetings restrictions imposed by the Covid-19 pandemic, all interviews were conducted virtually via video conferencing. We obtained the participant’s written consent to participate, via email, in advance of the interview. Participants were sent the interview questions in advance of the interviews. Verbal consent for the interview to be video recorded was obtained at the beginning of the interview. Interviews were conducted by between August and October 2020 and recruitment continued until no new insights emerged. Participants received $100.00 gift cards as compensation for their time.
Data analysis
The transcripts were transcribed verbatim and analyzed using thematic analysis [27]. Initially during the familiarization phase, one researcher (XX) read and coded the transcripts and identified initial a priori and emerging codes. To ensure inter-rater reliability, three research team members (XX, XX, XX) independently coded a sample of interview transcripts, which were compared against the first team member’s coding of the transcripts. Discrepancies were resolved through consultation with the team. The team developed and refined the codebook iteratively by re-coding and refining a priori and emerging themes. The transcripts and codes were entered into NVivo 10 software, a qualitative analysis package (QSR International, 2020). The number of instances was tabulated to confirm dominant themes. Code saturation was reached when no new codes were identified across all transcripts.
Findings
Sample characteristics
Thematic saturation was reached with 17 interviews. The characteristics of interventions, participants and LVC are in Table 1. The participants represented Choosing Wisely Canada implementation efforts in four provinces across Canada from hospital (n = 15) and primary care (n = 2) settings and addressed a variety of LVC (see Table 1).
Table 1
Location | Ontario | | 12 |
| Newfoundland | | 3 |
| Saskatchewan | | 1 |
| Nova Scotia | | 1 |
Setting | Hospital | | 15 |
| Primary Care | | 2 |
Role in Choosing Wisely Canada Intervention | Team Lead | | 12 |
Physician Champion | | 2 |
Team Member | | 3 |
Low-value care | Prescribing | Opioids | 3 |
| | Antibiotics | 2 |
| | PPIs | 1 |
| Laboratory testing | | 5 |
| Pre-Operative testing | | 2 |
| Blood transfusions | | 2 |
| Imaging | | 1 |
| Catheter use | | 1 |
Key findings
Our analysis of the data identified three major themes with subthemes relating to the research questions: drivers of LVC; the magnitude of the problem of LVC; and unique influences on de-implementation processes (see Fig. 1.).
Theme 1: Drivers of LVC
The first theme describes participant’s perspectives on the drivers of LVC. The majority of participants provided perspectives on factors deemed significant to the drivers of the targeted LVC. Participants noted the importance of understanding why the practice had reached inappropriate levels and also understanding what factors were sustaining the practice to better develop and implement interventions to reduce the LVC. Provider factors were prominently discussed, while patients were not identified as significant factors for the majority of LVC.
1.1 Provider factors
The most common factors discussed related to providers, such as habituation, years of practice, training, and a belief that ‘more is better’. A low-value practice being done through habit was identified as a significant factor in the sustainment of the practice. Some participants noted how habit extended to the institution and society.
I think a lot of the resistance was just related to people having their own style of practice, they've been doing [it] a certain way for so many years … probably the strongest resistance to this project was the force of habit. [P10]
In addition to the practice being ingrained for the provider, the practice being perceived as ‘status quo’ and supported by the system and patient expectations were identified as factors:
…the biggest barrier is culture, that this is the way you've practiced for a long time and this is the way that the population is expecting practice. It's the combination of demand from the population and a system where it's easier to meet the demand than push back on the demand. [P12]
Many participants pointed to years of practice or training, either medical school or in the institution, as additional ways the LVC was sustained:
There was a lot of people who were very resistant as well. So, there's quite a bit who had got quite defensive and put their backs up. And usually, it was kind of the older physicians that I would say they were kind of set in their ways. [P17]
Participants noted that providers endeavour to provide optimal care for their patients and this can often motivate more care than is necessary. Participants reported that providers are motived to continue LVC because of concerns about misdiagnosis, reputation and a desire to meet expectations.
There was a culture of wanting to be very thorough and doing a lot of testing in order to demonstrate that you were really keeping an eye on things and being expansive in your differential diagnosis. So, there's that aspect of wanting to impress. [P03]
1.2: Patient factors
The majority of the LVC in this study were not patient-facing, meaning that patients demands or expectations or clinician perceptions of these patient-drivers of LVC, did not exert significant influence on the sustainment of the practice. Duplicative lab testing and imaging, unnecessary blood transfusion volumes, indwelling catheters, and antibiotics in the ICU are some of the practices which were not patient-facing and therefore were not influenced by patient expectations and demands.
Theme 2: The magnitude of the problem of LVC
The second theme describes the significance of understanding the magnitude of the problem of LVC through the concepts of harm, resources and prevalence. In the analysis, we identified that harm occurs on multiple levels. Participants discussed the significance of harm and how the recognition of this harm motivated the change. Resources were also a significant factor that motivated the decision to reduce the LVC. These Choosing Wisely Canada implementation efforts took many aspects of resources – from time to human resources, to financial – into consideration when deciding to reduce the LVC. Finally, the prevalence was an important yet complex factor. Project teams recognized that the level of the LVC was inappropriate but were often challenged to identify specifically how prevalent the practice was.
2.1 Harm
Almost all of the participants discussed the harm from the LVC as an important factor motivating efforts to reduce the practice. Harm can come from the actual performance of the LVC, from the potential or common downstream effects of the LVC, from longer-term effects on patients, or from downstream harm to population health. The physical harm to the patient from doing the actual procedure or practice was predominantly reported as the least significant harm. This direct harm, e.g., an additional blood draw or excessive radiation, was deemed minimal compared to other types of harm from the LVC:
…although x-rays have low exposure, the dose [of radiation from] a rib x-ray which requires several views of both ribs, definitely is a concern for harm. That's the main, I would say, patient harm. [P02]
Potential and common physical harm to the patient resulting from the LVC, such as infections, antimicrobial resistance or overtreatment were commonly discussed:
…causing an infection and that infection can spread and it can infect your orthopedic implant and that can be a pretty horrendous complication if that happens. But even short of that, just having a UTI is a problem. Patients can get disseminated sepsis from that. So that's a big problem and then just delirium as well, from having the catheter, from having a UTI. [P16]
Over-testing and overtreatment as a result of the LVC was also discussed:
When urine cultures are ordered incorrectly, they [can] lead to antimicrobial overuse… because we're doing the testing inappropriately, you're going to get a 15 to 50% positive rate of positive bacteria, which will lead to treatment and that has no benefit. [P02]
Addiction, overdoses and infections were discussed as potential or common downstream harm to patients:
There is still a significant portion of patients who overdose on medications like Hydromorphone or prescribed Fentanyl patches or Morphine for that matter. So, that's still a significant problem. [P04]
The impact of the LVC on population health was an important consideration for tackling the issue at most participants’ institutions:
…if we start overusing antibiotics the bacteria become resistance, then you're going to have troubles down the road where people actually need these bacteria antibiotics, and the antibiotics are not going to work. They're not going to be lifesaving. [P17]
2.2 Resources
The resource aspect of LVC was discussed by all participants. Resources belong to a broad category that encompasses patient and provider time, medical equipment, and supplies. A few participants stated that for their Choosing Wisely Canada effort, harm was not the primary driver, but rather wasted resources. One participant detailed the multiple aspects of wasted resources presented by the LVC:
… you're using more [blood] products and the products are valuable products [and] are not always available. You're using more lab time in doing the cross match and the issuing of the units, you are using more nursing time spending the time transfusing the unit. You are using more tubing system because each system has to have tube as well and you're using the patient's time, sitting there and receiving the transfusion. Giving an unneeded intervention is a huge waste of resources. [P15]
The waste of resources, not physical harm to patients, was a significant motivator for some of the initiatives:
… it would be more harm in the sense of wasted resources on the system, more than actual patient harm I would say. Because it's really hard to see, the idea of unnecessary testing is an important one, but because there's very little patient harm coming from it, it's hard to sort of make an argument for it. [P11]
…one of our challenges in MRI is we have a wait list. And so those cases should be for indications that require MRI because there's no other diagnostic. So, we really just don't want to be using up the time on stuff that's not going to change management [of care]. [P09]
Some participants discussed not only the immediate resources wasted, but also downstream waste produced and the burden to the healthcare system:
“We have a very high opioid overdose rate, [x] times the rest of the province and then our hospitalization and emergency department visits were also about [x] times [the] rest of the province. So, these patients, they certainly can take up a lot of resources. These patients go to walk in clinics, the emergency department multiple times a month. They're admitted for months at a time with infectious complications from injection IV drug use, endocarditis, all these sorts of things. So, even the prevention of one or two of these patients I think has a significant impact to resources of the healthcare system. [P04]
The volume of tests and procedures was taken into account when assessing the impact on resources:
…the cost of doing the test is quite low, maybe it's about two bucks a test. But the quantity of testing is so high that it translates into a substantial amount of spending. [P11]
2.3 Prevalence
In the context of the magnitude of the problem, the prevalence of the LVC was discussed by the majority of participants. Often participants knew how often the practice was being performed, but not how often the practice was being performed inappropriately. Some project teams collected data on practice rates pre-intervention while others started the intervention, with the soft goal of reducing inappropriate practice, without baseline data.
So, I would say about three-quarters of them were getting blood work. About two-thirds were getting ECGs and about 25% were getting chest x-rays. And these are numbers that could all essentially go to zero because we're already talking about the population that doesn't need them. We're talking about low-risk patients getting low-risk surgery. [P08]
Theme 3: Unique influences on de-implementation processes
The third major theme identified provides insights on participant perspectives on the unique influences on de-implementation processes. The interview guide walked participants through the implementation process of the interventions to reduce the LVC. Participants discussed various aspects of the implementation process, from planning to implementation, to monitoring and evaluation. Participants were asked if they used theory to develop their strategy, identify barriers and facilitators, or select intervention strategies. While a number of initiatives were quality improvement projects, none of the participants reported using theory to inform or guide the initiative. The subthemes described in this section highlight some unique aspects of de-implementation processes.
3.1 Choosing Wisely as a change influencer
Participants discussed many aspects of how the Choosing Wisely Canada campaign supported de-implementation efforts. They discussed that Choosing Wisely Canada is respected, well known in the medical community, and had done much to increase awareness about LVC and the benefits of reducing it. The impact of Choosing Wisely campaigns was also perceived to generally influence culture and support providers to question existing practices:
Choosing Wisely was really the catalyst of de-implementation. It was an awareness. It was a lot of education here. [P05]