Treatment of colorectal cancer (CRC) involves surgical removal of the tumor and mesenteric lymph nodes (LN). Subsequent decisions about the need for additional treatment are partially based on whether pathologic findings demonstrate the presence of cancer in the LNs, since adjuvant chemotherapy confers a 15% absolute survival benefit for those with LN metastases 1. Inadequate LN harvest has been associated with decreased survival, so several international cancer control agencies, including the National Cancer Institute, American Joint Committee on Cancer, and the International Union Against Cancer now recommend examination of a minimum of 12 LNs to optimize CRC staging 23.

We conducted a population-based study revealing that 27% of eligible CRC patients in the province of Ontario, Canada had an adequate LN assessment 4. Dual efforts were consequently undertaken to improve this practice. In 2004 our group launched a randomized study (RCT) to examine the effect of engaging local surgeon or pathologist opinion leaders at participating hospitals on LN staging practice compared with a visiting educational presentation delivered to surgeons and pathologists (ISRCTN56824239) 5. At the same time the provincial cancer agency (PCA) developed a series of CRC performance measures 6. In 2005 the PCA distributed regional performance data to hospital CEOs and Chiefs of Surgery, including compliance with CRC LN staging recommendations. One year later the PCA reported that 58% of eligible CRC surgeries had 12 or more lymph nodes reported, a large change from the 27% revealed by population-based study, but this varied considerably across regions (range 33% to 92%) so there remained a need for ongoing quality improvement. Then the RCT was concluded and found that LN staging practice had improved in both the control and experimental arms of the trial 7. There was no significant difference between the two arms in proportion of patients with 12 or more LN removed (64% control, 76% experimental). There was no consistent effect of academic status or hospital volume on LN retrieval rates. This study did not reveal how changes were achieved, or the factors influencing LN staging practice, information that could guide quality improvement efforts to further improve compliance.

The interventions employed by these initiatives, performance data, educational presentation and opinion leaders, represent distinct strategies for influencing care delivery. Public reporting of performance data has improved outcomes in some circumstances, for example, decreased mortality after cardiac surgery, but research suggests that such data has little effect on physician practice 891011. Managers are thought to be a better target for performance data because they wish to enhance their hospital's public image, particular if performance accountability agreements are in place 12. Even so, it appears that managers may use performance data only when working in an institution that fosters a culture of quality, which suggests that organizational factors may be just as important as the information delivery mechanism 13.

Numerous interventions directed at physicians can improve compliance with practice recommendations but not across all conditions or settings of care, again highlighting the importance of organizational context 14. For example, outreach visits involving educational presentations have been most useful for promoting prescribing practice 15. Although opinion leaders appear to reduce non-compliance with desired practice, their effect appears comparable to that of educational visiting presentations or distribution of performance data, and studies have inconsistently described the actions of opinion leaders so the way they exert an effect in different contexts is not known 16. Furthermore, the impact of these interventions has been largely examined under rigorous experimental conditions, which measure specific outcomes, but provide little insight on the circumstances underlying those findings, or guidance on how to best implement those strategies in the real world.

Several factors are known to enable or hamper the adoption of a new practice including the mechanism by which information about the practice is delivered, attributes of the practice, physician characteristics, and elements of the practice environment 171819. In particular, the importance of organizational capacity for facilitating change is increasingly evident 2021222324. A meta-analysis of data from 81 studies examining interventions designed to increase the use of screening services for colon, breast and cervical cancer by physicians found that organizational change strategies influenced use of services more than reminders, feedback of performance data, education, financial incentives, legislative action, or mass media campaign 25. It is thought that organizational capacity includes the ability to identify, interpret, share and apply new knowledge or technology 26. However, it is not known what organizational structures and processes optimize the adoption of different innovative practices.

Compliance with LN staging recommendations has improved, but inconsistently. The way in which different, theoretically-promising interventions, including performance data, educational presentation and opinion leaders may have had an effect on LN staging practice is unknown. It is possible that some organizations responded differently because they better processed and applied the information that was delivered through these interventions. No studies have investigated the factors that influence CRC LN staging practice. To reveal issues amenable to quality improvement we examined how interventions used to deliver LN staging information, and the characteristics of LN staging practice, physicians, and their environment contributed to changes in practice. The findings could be used to optimize the design of ongoing quality improvement efforts. They would also increase our understanding of how to better equip organizations to support changes in practice, and apply change interventions in a more effective manner.

Interviews were conducted with 17 surgeons and nine pathologists whose characteristics are shown in Table 3. This included five surgeons and three pathologists from three of 11 eligible hospitals in three different health regions exposed to the educational presentation-only; six surgeons and three pathologists from five of nine eligible hospitals in three different health regions exposed to the opinion leader strategy; and six surgeons and three pathologists from three of ten eligible hospitals in three different health regions exposed to performance data. Hence, opinions were gathered from individuals at several different hospitals representing each of three information delivery interventions and seven different health regions. The way in which characteristics of LN staging practice, physicians and their organizations influenced response to educational presentation, opinion leader and performance data interventions is discussed here with representative quotes identified by an anonymous code.

The purpose of this study was to examine how various factors influenced the outcome of different interventions implemented to increase compliance with LN staging recommendations, identify how this practice could be further improved, and generate knowledge on the structures and processes required in health delivery organizations to better adopt innovative clinical practices. No studies had qualitatively investigated this practice and, in taking an exploratory approach, several actionable issues were identified for practice and ongoing research.

Improvements in LN staging practice may have occurred largely due to educational presentations that created awareness about the new recommendations, and self-initiated changes undertaken by pathologists. There was little awareness of other interventions, so executives and managers that received performance data may not have shared this with staff, and opinion leaders engaged to promote compliance with the new practice may not have fulfilled their intended roles. Barriers to change that are potentially amenable to quality improvement included perceptions about the practice (lack of awareness of evidence for the need to examine a minimum of 12 LNs) and associated responsibilities (blaming other profession), technical issues (increased pathology effort, need for pathology assistants, better clearing solutions and laboratory facilities), and a lack of organizational support for multidisciplinary interaction (little formal or informal communication within and between surgical and pathology departments) or quality improvement (no leadership for change or capacity for monitoring).

Most surgeons in this study indicated they were not compelled to change their surgical technique but both surgeons and pathologists acknowledged that type of cancer and tumor location made LN retrieval challenging, and that the amount of tissue collected partially determined whether a minimum of 12 LN could be assessed. This study was not designed to examine individual lymph node recovery rates of participating surgeons, but future research might investigate whether and how surgical outcomes could be improved for cases with challenging tumor locations, and whether surgical decision-making or technical skill influence tissue retrieval 31. Surgical judgment, largely based on personal assessments about technical ability, knowledge and experience, may not be amenable to modification using traditional continuing education due to a limited capacity of physicians to self-assess. Indeed, a number of studies have found sub-optimal performance in physicians who were the least skilled but the most confident in their technical ability 32. A Cochrane review of 118 trials found that audit and feedback can be effective in improving clinical practice when individuals are able to compare their own performance with that of peers, data is provided periodically on an ongoing basis, and is delivered verbally or by senior colleagues 33. Therefore, if it were found that surgical lymph node retrieval warranted improvement, individualized, comparative performance feedback could raise self-awareness of potential deficits in technical ability, and perhaps prompt alternative surgical approaches.

Participants said that more opportunities for fostering interaction between surgeons and pathologists must be created and supported. There is growing evidence that multidisciplinary collaboration improves health care team effectiveness and patient outcomes 34. One option to consider is the tumor board or multidisciplinary cancer conference (MCC). These are regularly scheduled multidisciplinary meetings to prospectively review individual cancer patients and formulate appropriate management plans 35. Several studies suggest that MCCs can improve compliance with practice recommendations but more research is required to understand how MCCs can be most effectively implemented 3637383940.

Organizations monitoring, or attempting to influence performance cannot rely on distribution targets such as executives or clinical opinion leaders for communication of information to those engaged in the practice, or to implement change initiatives. Instead, performance reporting should be coupled with incentives for change. At this time the impact of pay-for-performance is unclear, and our participants were concerned about its implications 41. They also requested individualized versus regional performance data. Research has shown that response to performance data is improved when the intent is explicitly non-punitive, and individual performance is compared to that of peers 4243. Additional research is necessary to understand alternative non-financial rewards. We learned that CRC staging based on 12 LNs is not possible in all patients due to physiologic characteristics, and nature and location of the cancer. Organizations monitoring LN staging must therefore ensure that performance data is adjusted for these factors, otherwise compliance could be under-reported.

However, even when coupled with appropriate incentives, interventions are likely to have little impact when the target lacks a receptive infrastructure. Institutional capacity for change may be a strong predictor for the adoption of innovations, and this includes the actions of managers 2021222324. For example, in American hospitals found to be high-performing for use of beta-blockers after acute myocardial infarction, managers ensured that resources were available to support changes, and actively promoted interdepartmental collaboration 44. A survey of 1,784 American hospitals found that high involvement of managers in quality improvement significantly decreased inpatient mortality for several conditions [45]. Further investigations are required to describe and evaluate the structures and processes that contribute to quality improvement capacity.

Several limitations are inherent in the exploratory design of this study. First, the expressed opinions are particular to the involved participants, and their geographic and health system environment, and therefore may not be applicable to health professionals in other settings. However, such research is meant to generate hypotheses that can be tested through further descriptive research performed elsewhere. Second, recruitment in qualitative research typically involves a small number of participants that are not randomly selected, potentially biasing data collection. To mitigate this concern and optimize validity, we used purposive sampling so that a variety of possible opinions were collected from individuals representing several different hospitals within each intervention group and across different health regions. Furthermore, we achieved informational redundancy, both by profession and intervention group, underscoring the validity of our findings. Still, studies involving larger samples of hospitals and health professionals should be conducted to confirm the influence of opportunities for collegial interaction and organizational capacity for quality improvement on LN staging practice, as well as other clinical innovations.

Use of an exploratory approach provided an in-depth view of the way that numerous factors amenable to quality improvement influenced the adoption of new CRC LN staging recommendations. In particular, lack of organizational infrastructure to coordinate and monitor changes in practice, and support multidisciplinary interaction may have limited the response to CRC LN staging recommendations. Individualized rather than regional or punitive performance data, coupled with increased organizational capacity for change may stimulate greater surgical and organizational response to quality improvement. Descriptive or experimental studies are needed to test these hypotheses.

The author(s) declare that they have no competing interests.

ARG obtained funding for the study, conceived and designed the study, acquired data, analyzed and interpreted data, and prepared the manuscript. FCW contributed to design of the study, assisted with data analysis and contributed to preparation of the manuscript. MAK contributed to design of the study, assisted with interpretation of the data, contributed to preparation of the manuscript. AJS assisted with interpretation of the data and contributed to preparation of the manuscript. All authors read and approved the final manuscript.