The 310 studies originated from 38 countries, with almost half conducted in the USA (n = 136). The most frequent type of LVC was non-indicated antibiotics (n = 84), followed by potentially inappropriate medication for the elderly (n = 73), imaging (n = 41), and lab tests (n = 31). The most common study setting was hospitals (n = 153), followed by primary care (n = 87).
Of the 310 studies, 279 were based on quantitative methods, 25 were mixed methods, and 10 were qualitative. The most common study design was a pre–post study design (n = 147), followed by quasi-experimental study design (n = 66) and randomized controlled trial (n = 39). The most frequent type of evaluation was efficacy/effectiveness (n = 260), followed by process evaluation (n = 21). Of the 310 studies, 217 used multicomponent strategies to reduce LVC, and 93 used single-component strategies.
Identified de-implementation strategies
The inductive coding yielded 71 unique strategies, of which 62 could be mapped onto ERIC strategies. A total of 36 of the 73 ERIC strategies were covered. Four of the identified strategies could be mapped onto one of the additional strategies suggested by Perry et al.: assess and redesign workflow (36). Five of the identified strategies could not be mapped onto the ERIC compilation or the suggested additions from Perry et al. (36).
The de-implementation strategies used most commonly were related to the ERIC categories training and education of stakeholders, use of evaluative and iterative strategies, and support of clinicians (Table 3).
Table 3
Number and percentages of the identified de-implementation strategies.
Categories of Strategies in ERIC | Strategy in ERIC | Number of studies | Percentage of total no. of studies |
Train and educate stakeholders | Conduct ongoing training | 0 | 0% |
Provide ongoing consultation | 0 | 0% |
Develop educational materials | 106 | 34% |
Make training dynamic | 85 | 27% |
Distribute educational materials | 59 | 19% |
Use train-the-trainer strategies | 3 | 1% |
Conduct educational meetings | 64 | 21% |
Conduct educational outreach visits | 30 | 10% |
Create a learning collaborative | 3 | 1% |
Shadow other experts | 0 | 0% |
Work with educational institutions | 0 | 0% |
Sum strategies | 350* | |
Use evaluative and iterative strategies | Assess for readiness and identify barriers and facilitators | 10 | 3% |
Audit and provide feedback | 89 | 29% |
Purposively reexamine the implementation | 0 | 0% |
Develop and implement tools for quality monitoring | 3 | 1% |
Develop and organize quality monitoring systems | 86 | 28% |
Develop a formal implementation blueprint | 0 | 0% |
Conduct local needs assessment | 10 | 3% |
Stage implementation scale up | 2 | 1% |
Obtain and use patients/consumers and family feedback | 0 | 0% |
Conduct cyclical small tests of change | 6 | 2% |
Sum strategies | 206 | |
Support clinicians | Facilitate relay of clinical data to providers | 3 | 1% |
Remind clinicians | 92 | 30% |
Develop resource sharing agreements | 0 | 0% |
Revise professional roles | 2 | 1% |
Create new clinical teams | 3 | 1% |
Sum strategies | 100 | |
Develop stakeholder interrelationships | Identify and prepare champions | 15 | 5% |
Organize clinician implementation team meetings | 1 | 1% |
Recruit, designate, and train for leadership | 0 | 0% |
Inform local opinion leaders | 0 | 0% |
Build a coalition | 0 | 0% |
Obtain formal commitment | 9 | 3% |
Identify early adopters | 0 | 0% |
Conduct local consensus discussions | 1 | 1% |
Capture and share local knowledge | 0 | 0% |
Use advisory boards and workgroups | 28 | 9% |
Use an implementation advisor | 0 | 0% |
Model and simulate change | 0 | 0% |
Visit other sites | 0 | 0% |
Involve executive boards | 0 | 0% |
Develop an implementation glossary | 0 | 0% |
Develop academic partnerships | 0 | 0% |
Promote network weaving | 0 | 0% |
Sum strategies | 54 | |
Change infrastructure | Mandate change | 2 | 1% |
Change record systems | 1 | 1% |
Change physical structure and equipment | 38 | 12% |
Create or change credentialing and/or licensure standards | 0 | 0% |
Change service sites | 0 | 0% |
Change accreditation or membership requirements | 0 | 0% |
Start a dissemination organization | 1 | 1% |
Change liability laws | 0 | 0% |
Sum strategies | 40 | |
Utilize financial strategies | Fund and contract for clinical innovation | 0 | 0% |
Access new founding | 0 | 0% |
Place innovation on fee for service lists/formularies | 0 | 0% |
Alter incentive/allowance structures | 6 | 2% |
Make billing easier | 0 | 0% |
Alter patient/consumer fees | 3 | 1% |
Use other payment schemes | 4 | 1% |
Develop disincentives | 1 | 1% |
Use capitated payments | 0 | 0% |
Sum strategies | 14 | |
Adapt and tailor to context | Tailor strategies | 6 | 2% |
Promote adaptability | 0 | 0% |
Use data experts | 0 | 0% |
Use data warehousing techniques | 2 | 1% |
Sum strategies | 8 | |
Provide interactive assistance | Facilitation | 0 | 0% |
Provide local technical assistance | 0 | 0% |
Centralize technical assistance | 0 | 0% |
Provide clinical supervision | 5 | 2% |
Sum strategies | 5 | |
Engage consumers | Involve patients/consumers and family members | 0 | 0% |
Intervene with patients/consumers to enhance uptake and adherence | 0 | 0% |
Prepare patients/consumers to be active participants | 0 | 0% |
Increase demand | 0 | 0% |
Use mass media | 2 | 1% |
Sum strategies | 2 | |
Added strategy by Perry et al. | Assess and redesign workflow | 21 | 7% |
Create online learning communities | 0 | 0% |
Engage community resources | 0 | 0% |
Strategies not found | ACCOUNTABILITY TOOL | 22 | 7% |
FDA BLACK BOX WARNING | 1 | 1% |
POLICY AND REGULATIONS | 5 | 2% |
COMMUNICATION TOOL | 9 | 3% |
| INTERNATIONAL COLLABORATION | 1 | 1% |
* Number exceeds total number of strategies since many of the studies used multiple strategies. |
Starting with the most common category, Table 3 presents the identified de-implementation strategies for each of the nine categories in ERIC (35), the added strategies by Perry et al. (36), and strategies not reflected in any of the previous strategies. A table with all identified strategies in each study is provided in Additional file 4.
[Table 3]
Train and educate stakeholders
A majority of the identified de-implementation strategies were related to the category train and educate stakeholders. Develop educational materials (n = 106) was the most frequently used strategy in that category. This strategy was comprised of information tailored to various target audiences such as patients (e.g., (48, 49)) and practitioners (e.g., (50, 51)). Make training dynamic (n = 85) was the second most common strategy. This strategy included various types of staff trainings with active participation from the participants, including case studies, handouts, and a pre- and post-education knowledge test (e.g., (52, 53)). Another frequently used strategy included distribute educational materials (n = 59), which consisted of a more passive distribution of guidelines to practitioners (e.g., (54, 55)).
Use evaluative and iterative strategies
Use of evaluative and iterative strategies was the second most common category of ERIC strategies. Within this category, audit and provide feedback (n = 89) was the most frequently used de-implementation strategy. It included various types of targets for the feedback such as individuals (e.g., (51, 56) and teams (e.g., (57, 58)), as well as feedback targeted to high prescribers only (e.g., (59)) and combined with social comparisons (e.g., (60)) or benchmark data (e.g., (61)). Another common strategy within this group was develop and organize a quality monitoring system (n = 86). This strategy consisted of monitoring systems electronically (e.g., (62, 63)) or via a pharmacist (e.g., (64, 65)) or peer (e.g., (66)). It also included feedback on the clinical outcomes of the reduced use of LVC (e.g., (67)).
Support clinicians
This category of ERIC strategies was the third most common. The majority of the identified de-implementation strategies in the category used the ERIC strategy remind clinicians (n = 92). This entailed digital (e.g., (68, 69)) or analog (e.g., (70, 71)) clinical decision support or other types of reminders such as stickers (e.g., (72)).
Develop stakeholder interrelationships
In this category of ERIC strategies, use advisory boards and workgroups (n = 28) was the most common de-implementation strategy. This category consisted of studies that had involved staff in planning the strategy (e.g., (73, 74)). Sixteen studies also used the strategy identify and prepare champions (e.g., (75, 76)), and 10 studies used obtain formal commitments (e.g., (77, 78)).
Change infrastructure
Within this category of ERIC strategies, the most frequently used strategy for de-implementation was change physical structure and equipment (n = 38). This encompassed changes in the ordering system for lab tests (e.g., (55, 79), changes in prescription process concerning medications (e.g., (80, 81)), facilitation of testing to only prescribe to patients with a certain test result (e.g., (82, 83)), facilitation of alternative practice (e.g., (84, 85)), and restricted availability of LVC practices (e.g., (86)).
Utilize financial strategies
In this category, alter incentive/allowance structure (n = 6) was the most common ERIC strategy. This strategy involved changing the level of reimbursement for LVC practices or the addition of criteria in the incentive system related to LVC use (e.g., (87). The category alter patient/consumer fees (n = 3) included both increased patient costs for LVC (e.g., (88)) and reduced patient costs for diagnostic tests that may hinder non-indicated antibiotic prescriptions (e.g.,(67)).
Adapt and tailor to context
In this category, tailor strategies (n = 6) was the most common ERIC strategy. The researchers used various methods to tailor the strategies to specific contexts, such as an educational strategy based on previously assessed knowledge among staff (e.g., (89)).
Provide interactive assistance
The only strategy that had been used for de-implementation in this category was provide clinical supervision (n = 5), which involved studies in which clinicians received supervision regarding when and how to use LVC practices and when other practices would be more beneficial for the patients. The supervision was either tailored to high prescribers (e.g., (90) or on demand for clinicians who requested it (e.g., (91)).
Engage consumers
The only strategy identified within this category was use mass media (n = 2). This strategy was represented by two studies in which an education campaign targeting the general population was conducted (92, 93).