- Email: [email protected]
Quality indicators in the management of bladder cancer: A modified Delphi study
Urologic Oncology: Seminars and Original Investigations ] (2016) ∎∎∎–∎∎∎
Original article
Quality indicators in the management of bladder cancer: A modified Delphi study Satya R. Khare, M.Sc.N., M.B.A.a, Armen Aprikian, M.D.b, Peter Black, M.D.c, Normand Blais, M.D.d, Chris Booth, M.D.e, Fadi Brimo, M.D.f, Joseph Chin, M.D.g, Peter Chung, M.D.h, Darrel Drachenberg, M.D.i, Libni Eapen, M.D.j, Adrian Fairey, M.D.k, Neil Fleshner, M.D.l,m, Yves Fradet, M.D.n, Geoffrey Gotto, M.D.o, Jonathan Izawa, M.D.g, Michael Jewett, M.D.l,m, Girish Kulkarni, M.D.l,m, Louis Lacombe, M.D.n, Ron Moore, M.D.k, Chris Morash, M.D.p, Scott North, M.D.q, Ricardo Rendon, M.D.r, Fred Saad, M.D.s, Bobby Shayegan, M.D.t, Robert Siemens, M.D.e,u, Alan So, M.D.c, Srikala S. Sridhar, M.D.v, Samer L. Traboulsi, M.D.b, Wassim Kassouf, M.D.b,* a Department of Family Medicine, McGill University, Montreal, Quebec, Canada Department of Urology, McGill University Health Centre, Montreal, Quebec, Canada c Department of Urology, University of British Columbia, Vancouver, British Columbia, Canada d Division of Medical Oncology, University of Montreal, Montreal, Quebec, Canada e Department of Oncology, Queen’s University, Kingston, Ontario, Canada f Department of Pathology, McGill University Health Centre, Montreal, Quebec, Canada g Division of Urology, Western University, London, Ontario, Canada h Department of Radiation Oncology, Princess Margaret Cancer Centre and the University Health Network, University of Toronto, Toronto, Ontario, Canada i Division of Urology, University of Manitoba, Winnipeg, Manitoba, Canada j Division of Radiation Oncology, University of Ottawa, Ottawa, Ontario, Canada k Division of Urology, University of Alberta, Edmonton, Alberta, Canada l Departments of Surgery (Urology), Princess Margaret Cancer Centre and University Health Network, University of Toronto, Toronto, Ontario, Canada m Department of Surgical Oncology, Princess Margaret Cancer Centre and University Health Network, University of Toronto, Toronto, Ontario, Canada n Division of Urology, Laval University, Quebec City, Quebec, Canada o Division of Urology, University of Calgary, Calgary, Alberta, Canada p Division of Urology, University of Ottawa, Ottawa, Ontario, Canada q Division of Medical Oncology, University of Alberta, Edmonton, Alberta, Canada r Division of Urology, Dalhousie University, Halifax, Nova Scotia, Canada s Division of Urology, University of Montreal, Montreal, Quebec, Canada t Division of Urology, McMaster University, Hamilton, Ontario, Canada u Department of Urology, Queen’s University, Kingston, Ontario, Canada v Department of Medical Oncology, Princess Margaret Cancer Centre and the University Health Network, University of Toronto, Toronto, Ontario, Canada b
Received 17 September 2016; received in revised form 12 November 2016; accepted 6 December 2016
Abstract Background: Survival in patients with bladder cancer has only moderately improved over the past 2 decades. A potential reason for this is nonadherence to clinical guidelines and best practice, leading to wide variations in care. Common quality indicators (QIs) are needed to quantify adherence to best practice and provide data for benchmarking and quality improvement. Objective: To produce an evidence- and consensus-based list of QIs for the management of bladder cancer. Methods: A modified Delphi method was used to develop the indicator list. Candidate indicators were extracted from the literature and rated by a 27-member Canadian expert panel in several rounds until consensus was reached on the final list of indicators. In rounds with numeric ratings, a frequency analysis was performed. Corresponding author. Tel.: þ1-514-934-8246; fax: þ1-514-934-8297. E-mail address: [email protected] (W. Kassouf). *
http://dx.doi.org/10.1016/j.urolonc.2016.12.003 1078-1439/r 2016 Elsevier Inc. All rights reserved.
2
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Results: A total of 86 indicators were rated, 52 extracted from the literature and 34 suggested by the panel. After iterative rounds of ratings and discussion, a final list of 60 QIs spanning several disciplines and phases of the cancer care continuum was developed. Conclusions: This is the first study to comprehensively produce common QIs representing structure, process, and outcome measures in bladder cancer management. Though developed in Canada, these indicators can be used in other countries with slight modifications to track performance and improve care. r 2016 Elsevier Inc. All rights reserved.
Keywords: Bladder cancer; Quality improvement; Quality indicators; Outcome
1. Introduction Bladder cancer is the ninth most common cancer worldwide [1], and the fifth most common cancer in Canada [2]. It includes a wide spectrum of disease from low-risk non– muscle-invasive tumors to highly aggressive, and often lethal, muscle-invasive tumors. Accordingly, the management of bladder cancer is complex and covers a broad range of interventions. Despite advances in diagnosis and treatment modalities, only a moderate improvement in cancer-specific survival has been observed over the past 2 decades [3]. Furthermore, population-based data have shown lower patient survival than that of clinical trials or academic centers [4]. The reason for this survival discrepancy is multifactorial ranging from a lack of proper health care facilities to nonadherence to urological guidelines [5,6]. Nonadherence leads to marked variations in care with several studies showing a link to poorer outcomes [5,7]. In countries where care facilities and availability of treatment are less likely contributors to poor patient outcomes, standardized quality-of-care assessment can serve to track performance and subsequent effect on clinical outcomes across the health care system [8]. To that end, the objective of this study was to produce an evidence- and consensusbased list of quality indicators (QIs) spanning the bladder cancer care continuum with input from a multidisciplinary expert panel. The QIs can be used to quantify adherence to best practices and provide data for benchmarking and quality improvement. It is our hope that performance measurement against common QIs will encourage the advancement of practice standards, promote performance comparison across jurisdictions in efforts to improve care, and stimulate sharing of best practices. The ultimate goal is to improve clinical outcomes of patients with bladder cancer.
2. Material and methods QIs were developed using a modified Delphi approach. The Delphi method has been used in similar studies and involves iterative rounds with controlled feedback to gain consensus from a group of experts in a systematic manner [9,10]. The modification involves an in-person meeting during the consensus process; however, owing to logistical reasons, a video-conference was held instead. In short, a literature review resulted in a list of evidence-based QIs that
were compiled into a candidate indicator list. Several rating rounds were then conducted with a national expert panel to come to consensus on which indicators would constitute the final QI list. 2.1. Panel selection Expert panel members were selected using a nomination process. The medical advisory board of Bladder Cancer Canada and the executive board of the Canadian Urologic Oncology Group were asked to nominate experts in bladder cancer care across the spectrum of clinical disciplines. Potential members were e-mailed about the study and asked if they would be interested to participate. 2.2. Literature review An extensive literature review was performed by one of the authors (W.K.) and a hospital librarian using the databases MEDLINE and Embase. The search strategy used text words and relevant indexing to identify articles discussing bladder cancer QIs or appropriate care of patients with bladder cancer. The search was limited to the adult population and articles published within the past 15 years. Case reports, commentaries, and editorials were excluded. The full MEDLINE search strategy is shown in Appendix A and was applied to both databases with modifications to search terms as necessary. The full text of all relevant records was examined. QIs were extracted from records that directly discussed bladder cancer QIs, and extrapolated from records that discussed clinical care supporting an effect on outcome. These evidence-based QIs, called candidate indicators, were compiled into a list and reviewed against the literature for accuracy and comprehensiveness by another author (S.T.). The list went through a final review by 2 authors (S.K. and W.K.) to remove duplicates and clarify the wording. The final list of candidate indicators was then organized into a questionnaire format, grouped by the following domains: diagnosis, staging, treatment, prophylactic measures, organizational process, outcomes, follow-up, and case volume. 2.3. Rating round 1 The questionnaire was e-mailed to the panel and members were asked to rate the candidate indicators on a
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Likert scale ranging from 0 (extremely inappropriate) to 4 (extremely appropriate) according to the following criteria: Is the indicator relevant and important to quality of care? Is the indicator scientifically sound (valid and reliable)? Does the indicator allow for comparison across jurisdictions? Is the indicator amenable to action (is it under your control to change)? Members were also encouraged to add indicators that were felt to be important but not already included in the list. Responses were collated and a frequency analysis was performed. The indicator reached consensus for inclusion if Z70% of the panel rated the indicator as 3 or 4. As there is no standard inclusion cut-off across Delphi studies, we chose a higher percentage to reflect strong consensus among the panel members. Indicators that did not meet the inclusion cut-off but did not have strong ratings for exclusion (i.e., an average rating of 2 on the scale) were considered neutral. 2.4. Rating round 2 Indicators that were added by panel members in round 1 were compiled in a second questionnaire. This was e-mailed to the panel and members were asked to rate the indicators using the same Likert scale and rating criteria specified earlier. A frequency analysis was performed on the collated responses and again indicators either reached consensus for inclusion or were considered neutral. 2.5. Group discussion A report that reflected rating results from rounds 1 to 2 was sent to the panel for review, along with comments made by respondents on their respective questionnaires. Indicators that had a neutral rating were discussed among the panel members via video-conference until consensus on inclusion or exclusion was reached based on the majority response. The meeting was moderated by 2 of the authors (S.K. and W.K.). 2.6. Feasibility assessment All indicators that reached consensus for inclusion through the previous rounds and discussion were compiled in a third questionnaire. Members were asked to think of their own institution, as well as consult with information management personnel, and specify whether the indicators would be feasible for data collection. Responses were collated and a frequency analysis was performed. The indicator was considered feasible if Z70% of the expert panel noted “yes.” 2.7. QI finalization To facilitate a final review by the panel, a questionnaire with the list of indicators was sent to the panel and members were asked to choose only those indicators that, in their
3
opinion, needed further refinement in wording and timelines. Indicators for which Z70% of the panel did not recommend further refinement were accepted as is. The rest underwent minor modification and finalized in another conference call. The meeting was moderated by 2 of the authors (S.K. and W.K.). All analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC).
3. Results The expert panel was composed of 27 members: 4 medical oncologists, 2 radiation oncologists, 1 genitourinary pathologist, and 20 urologic oncologists. All nominated members participated in the study. A total of 42 relevant records were retrieved from the literature search. From review of these records, 52 candidate indicators were extracted and organized as follows: 5 for diagnosis, 8 for staging, 19 for treatment, 4 for prophylactic measures, 2 for organizational process, 6 for outcomes, 2 for follow-up, and 6 for case volume. In the first rating round, 31 indicators met consensus for inclusion, 21 indicators had a neutral rating, and 34 indicators were newly suggested by the panel. Of the newly suggested indicators (rating round 2), 14 met consensus for inclusion and 20 had a neutral rating. All members returned a completed questionnaire in both rounds. The video conference to discuss the neutral indicators had 20 members in attendance (2 medical oncologist, 2 radiation oncologists, and 16 urologic oncologists) with a total of 41 indicators discussed individually. Through consensus, 21 indicators were deleted and some indicators were reworded. Comments made by members in previous rounds were also addressed during the discussion. At the end of this meeting, 65 indicators remained (20 from the video conference, 31 from the first rating round, and 14 from the second rating round). For the feasibility assessment, 20 members returned completed questionnaires and 5 indicators were discarded after analysis. The final list included 60 QIs categorized as follows: 5 for diagnosis, 11 for staging, 24 for treatment, 4 for prophylactic measures, 1 for organizational process, 8 for outcomes, 3 for follow-up, and 4 for case volume. For finalizing the QIs, 21 members returned completed questionnaires and 10 indicators were chosen for further refinement after analysis. A conference call was held. Members who were not available e-mailed their comments to the moderators before the call. Indicators that were further refined were under the categories of diagnosis (4), staging (3), and treatment (3). The main point of deliberation on these indicators was related to timelines; mostly based on what was feasible across the country as per panel consensus. Overall, 3 of the 10 indicators were left unchanged.
4
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Table Evidence- and consensus-based quality indicators for bladder cancer. Phase of care
Indicator
Supporting evidencea
Diagnosis
For patients with gross hematuria, percentage who saw urology within 2 wk of the date of request for consultation For patients with microhematuria and 440 y of age, percentage who saw urology within 6 wk of the date of request for consultation For patients with gross hematuria, percentage who had complete workup within 4 wk of the urology visit For patients with asymptomatic microhematuria and 440 y of age, percentage who had complete workup within 12 wk of the urology visit Percentage of patients with documented performance status and comorbidities at time of bladder cancer diagnosis Percentage of newly diagnosed patients who had transurethral resection of the bladder tumor (TURBT) within 3 wk of diagnosis Percentage of newly diagnosed patients who received upper tract imaging within 1 mo peri TURBT Percentage of patients for whom there was surgical report documentation on visual completeness of TUR, depth of TUR, and examination under anesthesia (EUA) findings Percentage of pathology reports available within 1 wk of TURBT Percentage of newly diagnosed, high-risk patients who were informed of pathology within 2 wk of TURBT Percentage of patients with T1–2 tumors for whom pathology was reviewed by a genitourinary pathologist Percentage of pathology reports noting detrusor muscle in the pathologic specimen (indicating completeness of resection) For patients with T1 disease and whose pathology report noted absence of detrusor muscle, percentage who were restaged by TUR within 6 wk of initial resection Percentage of patients with muscle-invasive bladder cancer (MIBC) undergoing chest imaging (chest radiograph/chest computed tomography) as part of staging For radical cystectomy (RC), percentage of patients who have preoperative cTNM recorded For RC, percentage of patients who have pTNM recorded postoperatively For patients who were referred to medical or radiation oncology, percentage who were seen within 2 wk of the date of request for referral Percentage of patients undergoing concurrent chemoradiotherapy who have a complete TURBT before therapy Percentage of patients who initiated trimodality therapy (TMT) within 6 wk of last TURBT Percentage of patients undergoing radiotherapy who have image-guided radiation therapy For patients indicated for radiation, percentage who received chemosensitizer with radiation Percentage of patients who started first cycle of neoadjuvant chemotherapy (NAC) within 4 wk of date of request to consult medical oncology Percentage of patients without NAC who had RC within 6 wk of last TURBT Percentage of patients with NAC who had RC within 16 wk of initiation of NAC For patients who underwent cystectomy, percentage who had a preoperative consultation with the enterostomal therapist Percentage of patients with adequate lymph node dissection defined as 416 nodes For patients who underwent cystectomy and o70 y of age, percentage who received continent diversion Percentage of patients who had soft tissue positive margin at cystectomy For patients with MIBC, percent who received any definitive therapy (RC or TMT) Percentage of patients with MIBC who initiated curative intent therapy (NAC, TMT, or RC) within 6 wk of TURBT Percentage of eligible patients with MIBC on TURBT being referred to medical oncology preoperatively for consideration of NAC For patients with MIBC and normal estimated glomerular filtration rate, Eastern Cooperative Oncology Group (ECOG) of 0–1, and o80 y of age, percentage who received NAC For patients with MIBC and receiving NAC, percentage who completed a minimum of 3 cycles of cisplatin-based combination therapy For patients with MIBC and pT3–4 or pNþ, percentage for whom there was a consult to medical oncology postoperatively For patients with MIBC, percentage who were managed by a multidisciplinary team (e.g., a multidisciplinary bladder cancer clinic, presentation at a genitourinary tumor board, arranged consultations with medical and radiation oncology when appropriate) For patients with non–muscle-invasive bladder cancer (NMIBC), percentage who received postoperative instillation of intravesical chemotherapy
Retrospective
Staging
Treatment
Retrospective Expert opinion Expert opinion Suggested by expert panel Expert opinion Retrospective/guideline Recommendation/guideline Expert opinion Expert opinion Retrospective Meta-analysis Randomized controlled trial Suggested by expert panel Suggested by expert panel Suggested by expert panel Expert opinion Suggested by expert panel Expert opinion Suggested by expert panel Randomized controlled trial Recommendation Retrospective Retrospective Recommendation/guideline Systematic reviewb Retrospective Retrospective Randomized controlled trial Suggested by expert panel Suggested by expert panel Randomized controlled trial Randomized controlled trial Randomized controlled trial Expert opinion
Meta-analysis of randomized controlled trials
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
5
Table Continued Phase of care
Prophylactic measures (patients undergoing cystectomy) Organizational process and outcomes
Indicator
Supporting evidencea
Percentage of patients with high-risk NMIBC receiving bacillus Calmette-Guerin (BCG) Percentage of patients with high-risk NMIBC who initiated intravesical BCG within 4 wk of TURBT For patients with high-risk NMIBC, percentage who had intravesical BCG induction course with at least 1-y maintenance Percentage of patients with metastatic or unresectable bladder cancer receiving cisplatin-based systemic chemotherapy Percentage of patients who received intravenous antibiotics within 60 min before incision Percentage of patients who received pharmacologic thrombosis prophylaxis perioperatively Percentage of patients for whom a pneumatic compression device was used intraoperatively Percentage of patients who received pharmacologic thrombosis prophylaxis after discharge for a period of 4 wk Presence of quality assurance rounds for discussion of complications postcystectomy For patients undergoing TMT, percentage who had:
Randomized controlled trial Suggested by expert panel
(1) acute Z grade 3 Radiation Therapy Oncology Group (RTOG) toxicity (2) late (equal to or greater than 3 mo) Z grade 3 RTOG toxicity Percentage of patients who had a length-of-stay of r10 d postcystectomy Transfusion rate during hospital admission for post-RC patients who:
Follow-up
Case volume
(1) received NAC (2) did not receive NAC Percentage of patients who had a severe (Clavien grade III/IV) postoperative complication within 90 d of surgery (e.g., requiring surgical, endoscopic, or radiologic intervention OR requiring ICU management for life-threatening complications) Percentage of patients who were readmitted within 90 d of cystectomy Percentage of patients deceased within 90 d postcystectomy Percentage of patients with ureteroenteric anastomotic stricture within 1 y of RC Stage (at diagnosis)-specific 5-y recurrence-free, disease-specific and overall survival after RC or TMT Percentage of patients who had cystoscopy 3 mo after TURBT For patients who underwent cystectomy with ileal conduit, percentage who had follow-up with the enterostomal therapist within 1 mo postdischarge For patients with MIBC, percentage with upper tract imaging and metastatic work-up within 1 y of definitive therapy Annual hospital volume of TMT Annual hospital volume of RC Annual surgical volume of RC by each surgeon performing this procedure Annual hospital volume of neobladders
Randomized controlled trial Suggested by expert panel Randomized controlled trial Recommendation/guideline Recommendation/guideline Recommendation/guideline Suggested by expert panel
Recommendation/guideline Retrospective Suggested by expert panel
Recommendation
Prospective studies Prospective studies Suggested by expert panel Suggested by expert panel Prospective studies Expert opinion Suggested by expert panel Expert opinion Meta-analysis Retrospective Retrospective
ICU ¼ intensive care unit. a This column represents the highest level of evidence supporting the importance of the indicator, however, the associated timelines and specific wording were agreed upon by consensus. b A systematic review supported extended lymph node dissection by anatomic template, and retrospective studies supported a threshold of 16 nodes. The numeric measure was chosen for ease of data collection as opposed to anatomic.
The final QI list is shown in the Table and the phased progression of consensus among panel members is shown in the Fig.
4. Discussion Bladder cancer is a heterogeneous disease that requires care across many disciplines along the care continuum from diagnosis to palliation. Despite progress in treatment, survival in patients with bladder cancer has only moderately improved. A potential reason for this is nonadherence to clinical guidelines and best practice, leading to wide
variations in care and a possible link to inferior outcomes. In this study, an evidence- and consensus-based list of QIs for the management of bladder cancer was produced using a modified Delphi method. The indicators span practice disciplines (surgery, pathology, medical oncology, radiation oncology, and enterostomal therapy) as well as phases of the cancer care continuum. These indicators will be used to track adherence to best practices and performance benchmarks across the health care system, and inform quality improvement initiatives. Quality assessment in health care generally follows the Donabedian model of structure, process, and outcome; examining all 3 provides a good measure of health care
6
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Fig. Phased progression of indicator selection within the Delphi process.
quality [8,11]. Structure represents the capacity to provide quality care, process represents actual care delivery, and outcomes represent the effect of care on patients or population or both. The QIs in this study had a fair representation of all 3 categories. Indicators related to structure included hospital and provider volume, both of which are associated with oncologic outcomes [12–16]. Process indicators reflected care delivery in a timely manner and according to best practice (either based on evidence or expert consensus). Outcome indicators included postoperative complications, toxicity, and survival. In our QI list, the number of process indicators greatly exceeds the other 2 categories as these indicators provide more direct feedback on quality of care [8,17]. For example, a certain percentage of patients with high-risk non–muscle-invasive bladder cancer receiving bacillus Calmette-Guerin is a process indicator that provides immediate meaningful data that can be used to drive improvement initiatives, if the benchmark is not met. Conversely, an outcome indicator, such as overall survival at 5 years, involves a long lag time for feedback. Other outcome indicators in our list involve shorter timeframes, such as acute high-grade treatment toxicity, but this is the exception. Nonetheless, by including each category of the Donabedian model, as well as each phase of the cancer care trajectory, our QI list does not overemphasize one aspect of care over another. This is important as aspects of care that are measured tend to get more focus than nonmeasured aspects, which contradicts high-quality care [18]. The QI list presented here fills a notable gap in the literature on bladder cancer management. Clinical guidelines and recommendations have been described [4], however, studies reporting QIs are limited with only a few describing indicators that are either specific to 1 type of bladder cancer [19,20], or lack comprehensiveness [21]. This point was supported in the current study, as a candidate indicator list was constructed from a literature review but the panel added 34 indicators that were felt to be important to quality of care but not already included in the list. There are limitations in this study that deserve consideration. Some potential QIs did not reach consensus for inclusion in the final list, such as high-grade toxicity postchemotherapy and patient management
with enhanced recovery after surgery pathways [22]. Additionally, only a few QIs on systemic treatment of advanced disease made the final list. This could have been related to the 5 rounds of consensus building, which is more than what other Delphi studies report [9,10]. However, this did not compromise panel involvement and multiple rounds supported the integrity of the final QIs. Furthermore, outspoken panel members can sometimes exert more influence than others in the Delphi method [23], but there was anonymity in the questionnaire responses and conference calls were moderated by 2 people (one of whom was an independent researcher), thus mitigating this possibility. Also, related to the panel, there was no representation of psychosocial and palliative care, leading to an underrepresentation of QIs in these areas. However, QIs in both areas have been previously described for patients with cancer in general and could be added to the list [24–26]. That being said, because this study covered a wide spectrum of disciplines and phases of care, the number of QIs was high at 60 in total. A prioritization effort would be required to select a manageable number to start with, which would likely result in different countries tracking different QI subsets. This may impede direct international comparisons, however, if all jurisdictions within a nation collect data on the same subset, meaningful comparisons on a national level could be made. Although developed in the Canadian context, we believe the QIs have relevance in other countries interested in assessing adherence to best practice in bladder cancer management. The literature review performed in advance of the consensus process included all relevant literature on an international scale and was not specific to Canada or other universal health care systems. The QIs represent key structure and process measures that predict for quality outcomes irrespective of geographic location. Having said that, indicator modifications would likely be necessary, especially in relation to timelines associated with some indicators. In this study, there was debate and deliberation over appropriate timelines that were feasible in the Canadian health care system. Thus, it is expected that these would change in other locations. Nonetheless, this study is an important starting point in identifying standard QIs in bladder cancer. In the next phase of this initiative, a subset of QIs will be chosen, data definitions with numerators, denominators, and inclusion criteria will be created, and performance benchmarks will be set. The benchmarks will be taken into account, beyond quality of care, situations where adherence would not be expected (e.g., contraindication to chemotherapy, complications delaying, and treatment). Adherence to best practice will be quantified and compared across the country, as will variations in patient outcomes. This will inform improvement initiatives and promote collaboration across the country, with a goal of optimizing patient care and outcomes.
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
5. Conclusions Discrepancies in survival outcomes have been observed in bladder cancer and may be owing to nonadherence to best practices. Given the need for standard QIs to assess adherence to best practice, this is the first study to comprehensively produce a list of QIs using a modified Delphi method that incorporated evidence from existing literature as well as consensus among an expert group of clinicians. A total of 60 QIs representing structure, process, and outcome measures in bladder cancer management are presented. Although developed in Canada, the QIs can be used in other countries, with modifications as needed. These indicators can be used for quality care assessment and improvement in efforts to optimize clinical outcomes of patients with bladder cancer. Acknowledgments We would like to thank Bladder Cancer Canada for funding this work. The funding organization had no involvement in the study design, implementation, or writing of the manuscript. Appendix A. Supplementary material Supplementary data are available in the online version of this article at http://dx.doi.org/10.1016/j.urolonc.2016.12.003
References [1] Griffiths TR. Action on bladder cancer. Current perspectives in bladder cancer management. Int J Clin Pract 2013;65:435–48. [2] Canadian Cancer Society’s Advisory Committee on Cancer Statistics. Canadian cancer statistics 2015. Toronto, ON: Canadian Cancer Society, 2015;2015. [3] Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin 2016;66:7–30. [4] Kassouf W, Aprikian A, Black P, Kulkarni G, Izawa J, Eapen L, et al. Recommendations for the improvement of bladder cancer quality of care in Canada: a consensus document reviewed and endorsed by Bladder Cancer Canada (BCC), Canadian Urologic Oncology Group (CUOG), and Canadian Urological Association (CUA), December 2015. Can Urol Assoc J 2016;10:E46–80. [5] Chamie K, Saigal CS, Lai J, Hanley JM, Setodji CM, Konety BR, et al. Quality of care in patients with bladder cancer: a case report? Cancer 2012;118:1412–21. [6] Karl A, Adejoro O, Saigal C, Konety B, Urologic Diseases in America P. General adherence to guideline recommendations on initial diagnosis of bladder cancer in the United States and influencing factors. Clin Genitourin Cancer 2014;12:270–7.
7
[7] Morris DS, Weizer AZ, Ye Z, Dunn RL, Montie JE, Hollenbeck BK. Understanding bladder cancer death: tumor biology versus physician practice. Cancer 2009;115:1011–20. [8] Albert JM, Das P. Quality assessment in oncology. Int J Radiat Oncol Biol Phys 2012;83:773–81. [9] Darling G, Malthaner R, Dickie J, McKnight L, Nhan C, Hunter A, et al. Quality indicators for non-small cell lung cancer operations with use of a modified Delphi consensus process. Ann Thorac Surg 2014;98:183–90. [10] Wood L, Bjarnason GA, Black PC, Cagiannos I, Heng DYC, Kapoor A, et al. Using the Delphi technique to improve clinical outcomes through the development of quality indicators in renal cell carcinoma. J Oncol Pract 2013;9:262–7. [11] Donabedian A. The quality of care. How can it be assessed? J Am Med Assoc 1988;260:1743–8. [12] Casey RG, So AI. Bladder cancer: are provider volumes a suitable measure of quality of care? Nature Rev Urol 2011;8:361–3. [13] Goossens-Laan CA, Gooiker GA, van Gijn W, Post PN, Bosch JL, Kil PJ, et al. A systematic review and meta-analysis of the relationship between hospital/surgeon volume and outcome for radical cystectomy: an update for the ongoing debate. Eur Urol 2011;59:775–83. [14] Gore JL, Wright JL, Daratha KB, Roberts KP, Lin DW, Wessells H, et al. Hospital-level variation in the quality of urologic cancer surgery. Cancer 2012;118:987–96. [15] Nielsen ME, Mallin K, Weaver MA, Palis B, Stewart A, Winchester DP, et al. Association of hospital volume with conditional 90-day mortality after cystectomy: an analysis of the National Cancer Data Base. BJU Int 2014;114:46–55. [16] Siemens DR, Mackillop WJ, Peng Y, Berman D, Elharram A, Rhee J, et al. Processes of care and the impact of surgical volumes on cancerspecific survival: a population-based study in bladder cancer. Urology 2014;84:1049–57. [17] Mant J. Process versus outcome indicators in the assessment of quality of health care. Int J Qual Health Care 2001;13:475–80. [18] Organisation for Economic Co-operation and Development. Improving value in health care: measuring quality. Quality of Care Forum. Paris, France; 2010. [19] Cooperberg MR, Porter MP, Konety BR. Candidate quality of care indicators for localized bladder cancer. Urol Oncol 2009;27: 435–442. [20] Goossens-Laan CA, Kil PJ, Roukema JA, Bosch JL, De Vries J. Quality of care indicators for muscle-invasive bladder cancer. Urol Int 2011;86:11–8. [21] Montgomery JS, Miller DC, Weizer AZ. Quality indicators in the management of bladder cancer. J Natl Compr Canc Netw 2013; 11:492–500. [22] Tyson MD, Chang SS. Enhanced recovery pathways versus standard care after cystectomy: a meta-analysis of the effect on perioperative outcomes. Eur Urol 2016;70(6):995–1003. [23] Jones J, Hunter D. Consensus methods for medical and health services research. Br Med J 1995;311:376–80. [24] Earle CC. Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol 2003;21:1133–8. [25] Weissman DE, Meier DE. Operational features for hospital palliative care programs: consensus recommendations. J Palliat Med 2008; 11:1189–94. [26] Weissman DE, Meier DE. Center to advance palliative care inpatient unit operational metrics: consensus recommendations. J Palliat Med 2009;12:21–5.
Original article
Quality indicators in the management of bladder cancer: A modified Delphi study Satya R. Khare, M.Sc.N., M.B.A.a, Armen Aprikian, M.D.b, Peter Black, M.D.c, Normand Blais, M.D.d, Chris Booth, M.D.e, Fadi Brimo, M.D.f, Joseph Chin, M.D.g, Peter Chung, M.D.h, Darrel Drachenberg, M.D.i, Libni Eapen, M.D.j, Adrian Fairey, M.D.k, Neil Fleshner, M.D.l,m, Yves Fradet, M.D.n, Geoffrey Gotto, M.D.o, Jonathan Izawa, M.D.g, Michael Jewett, M.D.l,m, Girish Kulkarni, M.D.l,m, Louis Lacombe, M.D.n, Ron Moore, M.D.k, Chris Morash, M.D.p, Scott North, M.D.q, Ricardo Rendon, M.D.r, Fred Saad, M.D.s, Bobby Shayegan, M.D.t, Robert Siemens, M.D.e,u, Alan So, M.D.c, Srikala S. Sridhar, M.D.v, Samer L. Traboulsi, M.D.b, Wassim Kassouf, M.D.b,* a Department of Family Medicine, McGill University, Montreal, Quebec, Canada Department of Urology, McGill University Health Centre, Montreal, Quebec, Canada c Department of Urology, University of British Columbia, Vancouver, British Columbia, Canada d Division of Medical Oncology, University of Montreal, Montreal, Quebec, Canada e Department of Oncology, Queen’s University, Kingston, Ontario, Canada f Department of Pathology, McGill University Health Centre, Montreal, Quebec, Canada g Division of Urology, Western University, London, Ontario, Canada h Department of Radiation Oncology, Princess Margaret Cancer Centre and the University Health Network, University of Toronto, Toronto, Ontario, Canada i Division of Urology, University of Manitoba, Winnipeg, Manitoba, Canada j Division of Radiation Oncology, University of Ottawa, Ottawa, Ontario, Canada k Division of Urology, University of Alberta, Edmonton, Alberta, Canada l Departments of Surgery (Urology), Princess Margaret Cancer Centre and University Health Network, University of Toronto, Toronto, Ontario, Canada m Department of Surgical Oncology, Princess Margaret Cancer Centre and University Health Network, University of Toronto, Toronto, Ontario, Canada n Division of Urology, Laval University, Quebec City, Quebec, Canada o Division of Urology, University of Calgary, Calgary, Alberta, Canada p Division of Urology, University of Ottawa, Ottawa, Ontario, Canada q Division of Medical Oncology, University of Alberta, Edmonton, Alberta, Canada r Division of Urology, Dalhousie University, Halifax, Nova Scotia, Canada s Division of Urology, University of Montreal, Montreal, Quebec, Canada t Division of Urology, McMaster University, Hamilton, Ontario, Canada u Department of Urology, Queen’s University, Kingston, Ontario, Canada v Department of Medical Oncology, Princess Margaret Cancer Centre and the University Health Network, University of Toronto, Toronto, Ontario, Canada b
Received 17 September 2016; received in revised form 12 November 2016; accepted 6 December 2016
Abstract Background: Survival in patients with bladder cancer has only moderately improved over the past 2 decades. A potential reason for this is nonadherence to clinical guidelines and best practice, leading to wide variations in care. Common quality indicators (QIs) are needed to quantify adherence to best practice and provide data for benchmarking and quality improvement. Objective: To produce an evidence- and consensus-based list of QIs for the management of bladder cancer. Methods: A modified Delphi method was used to develop the indicator list. Candidate indicators were extracted from the literature and rated by a 27-member Canadian expert panel in several rounds until consensus was reached on the final list of indicators. In rounds with numeric ratings, a frequency analysis was performed. Corresponding author. Tel.: þ1-514-934-8246; fax: þ1-514-934-8297. E-mail address: [email protected] (W. Kassouf). *
http://dx.doi.org/10.1016/j.urolonc.2016.12.003 1078-1439/r 2016 Elsevier Inc. All rights reserved.
2
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Results: A total of 86 indicators were rated, 52 extracted from the literature and 34 suggested by the panel. After iterative rounds of ratings and discussion, a final list of 60 QIs spanning several disciplines and phases of the cancer care continuum was developed. Conclusions: This is the first study to comprehensively produce common QIs representing structure, process, and outcome measures in bladder cancer management. Though developed in Canada, these indicators can be used in other countries with slight modifications to track performance and improve care. r 2016 Elsevier Inc. All rights reserved.
Keywords: Bladder cancer; Quality improvement; Quality indicators; Outcome
1. Introduction Bladder cancer is the ninth most common cancer worldwide [1], and the fifth most common cancer in Canada [2]. It includes a wide spectrum of disease from low-risk non– muscle-invasive tumors to highly aggressive, and often lethal, muscle-invasive tumors. Accordingly, the management of bladder cancer is complex and covers a broad range of interventions. Despite advances in diagnosis and treatment modalities, only a moderate improvement in cancer-specific survival has been observed over the past 2 decades [3]. Furthermore, population-based data have shown lower patient survival than that of clinical trials or academic centers [4]. The reason for this survival discrepancy is multifactorial ranging from a lack of proper health care facilities to nonadherence to urological guidelines [5,6]. Nonadherence leads to marked variations in care with several studies showing a link to poorer outcomes [5,7]. In countries where care facilities and availability of treatment are less likely contributors to poor patient outcomes, standardized quality-of-care assessment can serve to track performance and subsequent effect on clinical outcomes across the health care system [8]. To that end, the objective of this study was to produce an evidence- and consensusbased list of quality indicators (QIs) spanning the bladder cancer care continuum with input from a multidisciplinary expert panel. The QIs can be used to quantify adherence to best practices and provide data for benchmarking and quality improvement. It is our hope that performance measurement against common QIs will encourage the advancement of practice standards, promote performance comparison across jurisdictions in efforts to improve care, and stimulate sharing of best practices. The ultimate goal is to improve clinical outcomes of patients with bladder cancer.
2. Material and methods QIs were developed using a modified Delphi approach. The Delphi method has been used in similar studies and involves iterative rounds with controlled feedback to gain consensus from a group of experts in a systematic manner [9,10]. The modification involves an in-person meeting during the consensus process; however, owing to logistical reasons, a video-conference was held instead. In short, a literature review resulted in a list of evidence-based QIs that
were compiled into a candidate indicator list. Several rating rounds were then conducted with a national expert panel to come to consensus on which indicators would constitute the final QI list. 2.1. Panel selection Expert panel members were selected using a nomination process. The medical advisory board of Bladder Cancer Canada and the executive board of the Canadian Urologic Oncology Group were asked to nominate experts in bladder cancer care across the spectrum of clinical disciplines. Potential members were e-mailed about the study and asked if they would be interested to participate. 2.2. Literature review An extensive literature review was performed by one of the authors (W.K.) and a hospital librarian using the databases MEDLINE and Embase. The search strategy used text words and relevant indexing to identify articles discussing bladder cancer QIs or appropriate care of patients with bladder cancer. The search was limited to the adult population and articles published within the past 15 years. Case reports, commentaries, and editorials were excluded. The full MEDLINE search strategy is shown in Appendix A and was applied to both databases with modifications to search terms as necessary. The full text of all relevant records was examined. QIs were extracted from records that directly discussed bladder cancer QIs, and extrapolated from records that discussed clinical care supporting an effect on outcome. These evidence-based QIs, called candidate indicators, were compiled into a list and reviewed against the literature for accuracy and comprehensiveness by another author (S.T.). The list went through a final review by 2 authors (S.K. and W.K.) to remove duplicates and clarify the wording. The final list of candidate indicators was then organized into a questionnaire format, grouped by the following domains: diagnosis, staging, treatment, prophylactic measures, organizational process, outcomes, follow-up, and case volume. 2.3. Rating round 1 The questionnaire was e-mailed to the panel and members were asked to rate the candidate indicators on a
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Likert scale ranging from 0 (extremely inappropriate) to 4 (extremely appropriate) according to the following criteria: Is the indicator relevant and important to quality of care? Is the indicator scientifically sound (valid and reliable)? Does the indicator allow for comparison across jurisdictions? Is the indicator amenable to action (is it under your control to change)? Members were also encouraged to add indicators that were felt to be important but not already included in the list. Responses were collated and a frequency analysis was performed. The indicator reached consensus for inclusion if Z70% of the panel rated the indicator as 3 or 4. As there is no standard inclusion cut-off across Delphi studies, we chose a higher percentage to reflect strong consensus among the panel members. Indicators that did not meet the inclusion cut-off but did not have strong ratings for exclusion (i.e., an average rating of 2 on the scale) were considered neutral. 2.4. Rating round 2 Indicators that were added by panel members in round 1 were compiled in a second questionnaire. This was e-mailed to the panel and members were asked to rate the indicators using the same Likert scale and rating criteria specified earlier. A frequency analysis was performed on the collated responses and again indicators either reached consensus for inclusion or were considered neutral. 2.5. Group discussion A report that reflected rating results from rounds 1 to 2 was sent to the panel for review, along with comments made by respondents on their respective questionnaires. Indicators that had a neutral rating were discussed among the panel members via video-conference until consensus on inclusion or exclusion was reached based on the majority response. The meeting was moderated by 2 of the authors (S.K. and W.K.). 2.6. Feasibility assessment All indicators that reached consensus for inclusion through the previous rounds and discussion were compiled in a third questionnaire. Members were asked to think of their own institution, as well as consult with information management personnel, and specify whether the indicators would be feasible for data collection. Responses were collated and a frequency analysis was performed. The indicator was considered feasible if Z70% of the expert panel noted “yes.” 2.7. QI finalization To facilitate a final review by the panel, a questionnaire with the list of indicators was sent to the panel and members were asked to choose only those indicators that, in their
3
opinion, needed further refinement in wording and timelines. Indicators for which Z70% of the panel did not recommend further refinement were accepted as is. The rest underwent minor modification and finalized in another conference call. The meeting was moderated by 2 of the authors (S.K. and W.K.). All analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC).
3. Results The expert panel was composed of 27 members: 4 medical oncologists, 2 radiation oncologists, 1 genitourinary pathologist, and 20 urologic oncologists. All nominated members participated in the study. A total of 42 relevant records were retrieved from the literature search. From review of these records, 52 candidate indicators were extracted and organized as follows: 5 for diagnosis, 8 for staging, 19 for treatment, 4 for prophylactic measures, 2 for organizational process, 6 for outcomes, 2 for follow-up, and 6 for case volume. In the first rating round, 31 indicators met consensus for inclusion, 21 indicators had a neutral rating, and 34 indicators were newly suggested by the panel. Of the newly suggested indicators (rating round 2), 14 met consensus for inclusion and 20 had a neutral rating. All members returned a completed questionnaire in both rounds. The video conference to discuss the neutral indicators had 20 members in attendance (2 medical oncologist, 2 radiation oncologists, and 16 urologic oncologists) with a total of 41 indicators discussed individually. Through consensus, 21 indicators were deleted and some indicators were reworded. Comments made by members in previous rounds were also addressed during the discussion. At the end of this meeting, 65 indicators remained (20 from the video conference, 31 from the first rating round, and 14 from the second rating round). For the feasibility assessment, 20 members returned completed questionnaires and 5 indicators were discarded after analysis. The final list included 60 QIs categorized as follows: 5 for diagnosis, 11 for staging, 24 for treatment, 4 for prophylactic measures, 1 for organizational process, 8 for outcomes, 3 for follow-up, and 4 for case volume. For finalizing the QIs, 21 members returned completed questionnaires and 10 indicators were chosen for further refinement after analysis. A conference call was held. Members who were not available e-mailed their comments to the moderators before the call. Indicators that were further refined were under the categories of diagnosis (4), staging (3), and treatment (3). The main point of deliberation on these indicators was related to timelines; mostly based on what was feasible across the country as per panel consensus. Overall, 3 of the 10 indicators were left unchanged.
4
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Table Evidence- and consensus-based quality indicators for bladder cancer. Phase of care
Indicator
Supporting evidencea
Diagnosis
For patients with gross hematuria, percentage who saw urology within 2 wk of the date of request for consultation For patients with microhematuria and 440 y of age, percentage who saw urology within 6 wk of the date of request for consultation For patients with gross hematuria, percentage who had complete workup within 4 wk of the urology visit For patients with asymptomatic microhematuria and 440 y of age, percentage who had complete workup within 12 wk of the urology visit Percentage of patients with documented performance status and comorbidities at time of bladder cancer diagnosis Percentage of newly diagnosed patients who had transurethral resection of the bladder tumor (TURBT) within 3 wk of diagnosis Percentage of newly diagnosed patients who received upper tract imaging within 1 mo peri TURBT Percentage of patients for whom there was surgical report documentation on visual completeness of TUR, depth of TUR, and examination under anesthesia (EUA) findings Percentage of pathology reports available within 1 wk of TURBT Percentage of newly diagnosed, high-risk patients who were informed of pathology within 2 wk of TURBT Percentage of patients with T1–2 tumors for whom pathology was reviewed by a genitourinary pathologist Percentage of pathology reports noting detrusor muscle in the pathologic specimen (indicating completeness of resection) For patients with T1 disease and whose pathology report noted absence of detrusor muscle, percentage who were restaged by TUR within 6 wk of initial resection Percentage of patients with muscle-invasive bladder cancer (MIBC) undergoing chest imaging (chest radiograph/chest computed tomography) as part of staging For radical cystectomy (RC), percentage of patients who have preoperative cTNM recorded For RC, percentage of patients who have pTNM recorded postoperatively For patients who were referred to medical or radiation oncology, percentage who were seen within 2 wk of the date of request for referral Percentage of patients undergoing concurrent chemoradiotherapy who have a complete TURBT before therapy Percentage of patients who initiated trimodality therapy (TMT) within 6 wk of last TURBT Percentage of patients undergoing radiotherapy who have image-guided radiation therapy For patients indicated for radiation, percentage who received chemosensitizer with radiation Percentage of patients who started first cycle of neoadjuvant chemotherapy (NAC) within 4 wk of date of request to consult medical oncology Percentage of patients without NAC who had RC within 6 wk of last TURBT Percentage of patients with NAC who had RC within 16 wk of initiation of NAC For patients who underwent cystectomy, percentage who had a preoperative consultation with the enterostomal therapist Percentage of patients with adequate lymph node dissection defined as 416 nodes For patients who underwent cystectomy and o70 y of age, percentage who received continent diversion Percentage of patients who had soft tissue positive margin at cystectomy For patients with MIBC, percent who received any definitive therapy (RC or TMT) Percentage of patients with MIBC who initiated curative intent therapy (NAC, TMT, or RC) within 6 wk of TURBT Percentage of eligible patients with MIBC on TURBT being referred to medical oncology preoperatively for consideration of NAC For patients with MIBC and normal estimated glomerular filtration rate, Eastern Cooperative Oncology Group (ECOG) of 0–1, and o80 y of age, percentage who received NAC For patients with MIBC and receiving NAC, percentage who completed a minimum of 3 cycles of cisplatin-based combination therapy For patients with MIBC and pT3–4 or pNþ, percentage for whom there was a consult to medical oncology postoperatively For patients with MIBC, percentage who were managed by a multidisciplinary team (e.g., a multidisciplinary bladder cancer clinic, presentation at a genitourinary tumor board, arranged consultations with medical and radiation oncology when appropriate) For patients with non–muscle-invasive bladder cancer (NMIBC), percentage who received postoperative instillation of intravesical chemotherapy
Retrospective
Staging
Treatment
Retrospective Expert opinion Expert opinion Suggested by expert panel Expert opinion Retrospective/guideline Recommendation/guideline Expert opinion Expert opinion Retrospective Meta-analysis Randomized controlled trial Suggested by expert panel Suggested by expert panel Suggested by expert panel Expert opinion Suggested by expert panel Expert opinion Suggested by expert panel Randomized controlled trial Recommendation Retrospective Retrospective Recommendation/guideline Systematic reviewb Retrospective Retrospective Randomized controlled trial Suggested by expert panel Suggested by expert panel Randomized controlled trial Randomized controlled trial Randomized controlled trial Expert opinion
Meta-analysis of randomized controlled trials
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
5
Table Continued Phase of care
Prophylactic measures (patients undergoing cystectomy) Organizational process and outcomes
Indicator
Supporting evidencea
Percentage of patients with high-risk NMIBC receiving bacillus Calmette-Guerin (BCG) Percentage of patients with high-risk NMIBC who initiated intravesical BCG within 4 wk of TURBT For patients with high-risk NMIBC, percentage who had intravesical BCG induction course with at least 1-y maintenance Percentage of patients with metastatic or unresectable bladder cancer receiving cisplatin-based systemic chemotherapy Percentage of patients who received intravenous antibiotics within 60 min before incision Percentage of patients who received pharmacologic thrombosis prophylaxis perioperatively Percentage of patients for whom a pneumatic compression device was used intraoperatively Percentage of patients who received pharmacologic thrombosis prophylaxis after discharge for a period of 4 wk Presence of quality assurance rounds for discussion of complications postcystectomy For patients undergoing TMT, percentage who had:
Randomized controlled trial Suggested by expert panel
(1) acute Z grade 3 Radiation Therapy Oncology Group (RTOG) toxicity (2) late (equal to or greater than 3 mo) Z grade 3 RTOG toxicity Percentage of patients who had a length-of-stay of r10 d postcystectomy Transfusion rate during hospital admission for post-RC patients who:
Follow-up
Case volume
(1) received NAC (2) did not receive NAC Percentage of patients who had a severe (Clavien grade III/IV) postoperative complication within 90 d of surgery (e.g., requiring surgical, endoscopic, or radiologic intervention OR requiring ICU management for life-threatening complications) Percentage of patients who were readmitted within 90 d of cystectomy Percentage of patients deceased within 90 d postcystectomy Percentage of patients with ureteroenteric anastomotic stricture within 1 y of RC Stage (at diagnosis)-specific 5-y recurrence-free, disease-specific and overall survival after RC or TMT Percentage of patients who had cystoscopy 3 mo after TURBT For patients who underwent cystectomy with ileal conduit, percentage who had follow-up with the enterostomal therapist within 1 mo postdischarge For patients with MIBC, percentage with upper tract imaging and metastatic work-up within 1 y of definitive therapy Annual hospital volume of TMT Annual hospital volume of RC Annual surgical volume of RC by each surgeon performing this procedure Annual hospital volume of neobladders
Randomized controlled trial Suggested by expert panel Randomized controlled trial Recommendation/guideline Recommendation/guideline Recommendation/guideline Suggested by expert panel
Recommendation/guideline Retrospective Suggested by expert panel
Recommendation
Prospective studies Prospective studies Suggested by expert panel Suggested by expert panel Prospective studies Expert opinion Suggested by expert panel Expert opinion Meta-analysis Retrospective Retrospective
ICU ¼ intensive care unit. a This column represents the highest level of evidence supporting the importance of the indicator, however, the associated timelines and specific wording were agreed upon by consensus. b A systematic review supported extended lymph node dissection by anatomic template, and retrospective studies supported a threshold of 16 nodes. The numeric measure was chosen for ease of data collection as opposed to anatomic.
The final QI list is shown in the Table and the phased progression of consensus among panel members is shown in the Fig.
4. Discussion Bladder cancer is a heterogeneous disease that requires care across many disciplines along the care continuum from diagnosis to palliation. Despite progress in treatment, survival in patients with bladder cancer has only moderately improved. A potential reason for this is nonadherence to clinical guidelines and best practice, leading to wide
variations in care and a possible link to inferior outcomes. In this study, an evidence- and consensus-based list of QIs for the management of bladder cancer was produced using a modified Delphi method. The indicators span practice disciplines (surgery, pathology, medical oncology, radiation oncology, and enterostomal therapy) as well as phases of the cancer care continuum. These indicators will be used to track adherence to best practices and performance benchmarks across the health care system, and inform quality improvement initiatives. Quality assessment in health care generally follows the Donabedian model of structure, process, and outcome; examining all 3 provides a good measure of health care
6
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
Fig. Phased progression of indicator selection within the Delphi process.
quality [8,11]. Structure represents the capacity to provide quality care, process represents actual care delivery, and outcomes represent the effect of care on patients or population or both. The QIs in this study had a fair representation of all 3 categories. Indicators related to structure included hospital and provider volume, both of which are associated with oncologic outcomes [12–16]. Process indicators reflected care delivery in a timely manner and according to best practice (either based on evidence or expert consensus). Outcome indicators included postoperative complications, toxicity, and survival. In our QI list, the number of process indicators greatly exceeds the other 2 categories as these indicators provide more direct feedback on quality of care [8,17]. For example, a certain percentage of patients with high-risk non–muscle-invasive bladder cancer receiving bacillus Calmette-Guerin is a process indicator that provides immediate meaningful data that can be used to drive improvement initiatives, if the benchmark is not met. Conversely, an outcome indicator, such as overall survival at 5 years, involves a long lag time for feedback. Other outcome indicators in our list involve shorter timeframes, such as acute high-grade treatment toxicity, but this is the exception. Nonetheless, by including each category of the Donabedian model, as well as each phase of the cancer care trajectory, our QI list does not overemphasize one aspect of care over another. This is important as aspects of care that are measured tend to get more focus than nonmeasured aspects, which contradicts high-quality care [18]. The QI list presented here fills a notable gap in the literature on bladder cancer management. Clinical guidelines and recommendations have been described [4], however, studies reporting QIs are limited with only a few describing indicators that are either specific to 1 type of bladder cancer [19,20], or lack comprehensiveness [21]. This point was supported in the current study, as a candidate indicator list was constructed from a literature review but the panel added 34 indicators that were felt to be important to quality of care but not already included in the list. There are limitations in this study that deserve consideration. Some potential QIs did not reach consensus for inclusion in the final list, such as high-grade toxicity postchemotherapy and patient management
with enhanced recovery after surgery pathways [22]. Additionally, only a few QIs on systemic treatment of advanced disease made the final list. This could have been related to the 5 rounds of consensus building, which is more than what other Delphi studies report [9,10]. However, this did not compromise panel involvement and multiple rounds supported the integrity of the final QIs. Furthermore, outspoken panel members can sometimes exert more influence than others in the Delphi method [23], but there was anonymity in the questionnaire responses and conference calls were moderated by 2 people (one of whom was an independent researcher), thus mitigating this possibility. Also, related to the panel, there was no representation of psychosocial and palliative care, leading to an underrepresentation of QIs in these areas. However, QIs in both areas have been previously described for patients with cancer in general and could be added to the list [24–26]. That being said, because this study covered a wide spectrum of disciplines and phases of care, the number of QIs was high at 60 in total. A prioritization effort would be required to select a manageable number to start with, which would likely result in different countries tracking different QI subsets. This may impede direct international comparisons, however, if all jurisdictions within a nation collect data on the same subset, meaningful comparisons on a national level could be made. Although developed in the Canadian context, we believe the QIs have relevance in other countries interested in assessing adherence to best practice in bladder cancer management. The literature review performed in advance of the consensus process included all relevant literature on an international scale and was not specific to Canada or other universal health care systems. The QIs represent key structure and process measures that predict for quality outcomes irrespective of geographic location. Having said that, indicator modifications would likely be necessary, especially in relation to timelines associated with some indicators. In this study, there was debate and deliberation over appropriate timelines that were feasible in the Canadian health care system. Thus, it is expected that these would change in other locations. Nonetheless, this study is an important starting point in identifying standard QIs in bladder cancer. In the next phase of this initiative, a subset of QIs will be chosen, data definitions with numerators, denominators, and inclusion criteria will be created, and performance benchmarks will be set. The benchmarks will be taken into account, beyond quality of care, situations where adherence would not be expected (e.g., contraindication to chemotherapy, complications delaying, and treatment). Adherence to best practice will be quantified and compared across the country, as will variations in patient outcomes. This will inform improvement initiatives and promote collaboration across the country, with a goal of optimizing patient care and outcomes.
S.R. Khare et al. / Urologic Oncology: Seminars and Original Investigations ] (2016) 1–7
5. Conclusions Discrepancies in survival outcomes have been observed in bladder cancer and may be owing to nonadherence to best practices. Given the need for standard QIs to assess adherence to best practice, this is the first study to comprehensively produce a list of QIs using a modified Delphi method that incorporated evidence from existing literature as well as consensus among an expert group of clinicians. A total of 60 QIs representing structure, process, and outcome measures in bladder cancer management are presented. Although developed in Canada, the QIs can be used in other countries, with modifications as needed. These indicators can be used for quality care assessment and improvement in efforts to optimize clinical outcomes of patients with bladder cancer. Acknowledgments We would like to thank Bladder Cancer Canada for funding this work. The funding organization had no involvement in the study design, implementation, or writing of the manuscript. Appendix A. Supplementary material Supplementary data are available in the online version of this article at http://dx.doi.org/10.1016/j.urolonc.2016.12.003
References [1] Griffiths TR. Action on bladder cancer. Current perspectives in bladder cancer management. Int J Clin Pract 2013;65:435–48. [2] Canadian Cancer Society’s Advisory Committee on Cancer Statistics. Canadian cancer statistics 2015. Toronto, ON: Canadian Cancer Society, 2015;2015. [3] Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin 2016;66:7–30. [4] Kassouf W, Aprikian A, Black P, Kulkarni G, Izawa J, Eapen L, et al. Recommendations for the improvement of bladder cancer quality of care in Canada: a consensus document reviewed and endorsed by Bladder Cancer Canada (BCC), Canadian Urologic Oncology Group (CUOG), and Canadian Urological Association (CUA), December 2015. Can Urol Assoc J 2016;10:E46–80. [5] Chamie K, Saigal CS, Lai J, Hanley JM, Setodji CM, Konety BR, et al. Quality of care in patients with bladder cancer: a case report? Cancer 2012;118:1412–21. [6] Karl A, Adejoro O, Saigal C, Konety B, Urologic Diseases in America P. General adherence to guideline recommendations on initial diagnosis of bladder cancer in the United States and influencing factors. Clin Genitourin Cancer 2014;12:270–7.
7
[7] Morris DS, Weizer AZ, Ye Z, Dunn RL, Montie JE, Hollenbeck BK. Understanding bladder cancer death: tumor biology versus physician practice. Cancer 2009;115:1011–20. [8] Albert JM, Das P. Quality assessment in oncology. Int J Radiat Oncol Biol Phys 2012;83:773–81. [9] Darling G, Malthaner R, Dickie J, McKnight L, Nhan C, Hunter A, et al. Quality indicators for non-small cell lung cancer operations with use of a modified Delphi consensus process. Ann Thorac Surg 2014;98:183–90. [10] Wood L, Bjarnason GA, Black PC, Cagiannos I, Heng DYC, Kapoor A, et al. Using the Delphi technique to improve clinical outcomes through the development of quality indicators in renal cell carcinoma. J Oncol Pract 2013;9:262–7. [11] Donabedian A. The quality of care. How can it be assessed? J Am Med Assoc 1988;260:1743–8. [12] Casey RG, So AI. Bladder cancer: are provider volumes a suitable measure of quality of care? Nature Rev Urol 2011;8:361–3. [13] Goossens-Laan CA, Gooiker GA, van Gijn W, Post PN, Bosch JL, Kil PJ, et al. A systematic review and meta-analysis of the relationship between hospital/surgeon volume and outcome for radical cystectomy: an update for the ongoing debate. Eur Urol 2011;59:775–83. [14] Gore JL, Wright JL, Daratha KB, Roberts KP, Lin DW, Wessells H, et al. Hospital-level variation in the quality of urologic cancer surgery. Cancer 2012;118:987–96. [15] Nielsen ME, Mallin K, Weaver MA, Palis B, Stewart A, Winchester DP, et al. Association of hospital volume with conditional 90-day mortality after cystectomy: an analysis of the National Cancer Data Base. BJU Int 2014;114:46–55. [16] Siemens DR, Mackillop WJ, Peng Y, Berman D, Elharram A, Rhee J, et al. Processes of care and the impact of surgical volumes on cancerspecific survival: a population-based study in bladder cancer. Urology 2014;84:1049–57. [17] Mant J. Process versus outcome indicators in the assessment of quality of health care. Int J Qual Health Care 2001;13:475–80. [18] Organisation for Economic Co-operation and Development. Improving value in health care: measuring quality. Quality of Care Forum. Paris, France; 2010. [19] Cooperberg MR, Porter MP, Konety BR. Candidate quality of care indicators for localized bladder cancer. Urol Oncol 2009;27: 435–442. [20] Goossens-Laan CA, Kil PJ, Roukema JA, Bosch JL, De Vries J. Quality of care indicators for muscle-invasive bladder cancer. Urol Int 2011;86:11–8. [21] Montgomery JS, Miller DC, Weizer AZ. Quality indicators in the management of bladder cancer. J Natl Compr Canc Netw 2013; 11:492–500. [22] Tyson MD, Chang SS. Enhanced recovery pathways versus standard care after cystectomy: a meta-analysis of the effect on perioperative outcomes. Eur Urol 2016;70(6):995–1003. [23] Jones J, Hunter D. Consensus methods for medical and health services research. Br Med J 1995;311:376–80. [24] Earle CC. Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol 2003;21:1133–8. [25] Weissman DE, Meier DE. Operational features for hospital palliative care programs: consensus recommendations. J Palliat Med 2008; 11:1189–94. [26] Weissman DE, Meier DE. Center to advance palliative care inpatient unit operational metrics: consensus recommendations. J Palliat Med 2009;12:21–5.