Abstract
Background:
Research examining the impact of different models of care on wait times for breast cancer surgery indicates that organized assessment can reduce wait times, but few researchers have explored different care approaches between care sites serving a mixture of urban and rural patients and breast cancer care outcomes, especially within the Canadian context. Therefore, we sought to explore whether wait times from mammogram to surgery differed between lean referral and traditional referral pathways and what areas of inefficiencies need to be addressed.
Methods:
We used a retrospective case review design and collected information on female patients (aged ≥ 19 yr) with stage 0-III breast cancer who were surgically treated between February 2016 and July 2019.
Results:
Patients referred in the traditional pathway of care (n = 208) had longer wait times than patients in the lean pathway of care (n = 248), with an average wait time difference of 9.3 days. For both care pathways, receiving a screening mammogram, living farther from the hospital, and receiving magnetic resonance imaging (MRI) increased wait times to surgery.
Conclusion:
Conducting the biopsy immediately after an abnormal mammogram, improving wait times for MRIs, and improving access for rural patients may be important areas of change-related focus. Shorter wait times to breast cancer surgery in the lean pathway support the advantage of a referral system whereby organized navigated systems coordinate all aspects of diagnosis and treatment.
Abstract
Contexte:
La recherche concernant l’impact de différents modèles de soins sur les temps d’attente en vue d’une chirurgie pour le cancer du sein révèle qu’un processus d’évaluation dûment coordonné permet de réduire les temps d’attente, mais peu de chercheurs ont exploré les diverses modalités utilisées par les centres qui desservent une population mixte urbaine et rurale, et leurs résultats sur le cancer du sein, particulièrement au Canada. Nous avons donc voulu vérifier si les délais qui séparent la mammographie de la chirurgie sont différents selon que l’aiguillage suit un processus allégé ou le processus habituel, et identifier les zones et les lacunes à corriger.
Méthodes:
Nous avons utilisé une revue rétrospective des dossiers et recueilli les données sur les patientes de 19 ans et plus atteintes d’un cancer du sein de stade 0–III traitées chirurgicalement entre février 2016 et juillet 2019.
Résultats:
Les temps d’attente ont été plus longs pour les patientes soumises à la trajectoire de soins habituelle (n = 208) que pour les patientes soumises au processus allégé (n = 248), avec une différence moyenne de 9,3 jours de temps d’attente. Pour les 2 trajectoires de soins, la mammographie de dépistage, l’éloignement de l’hôpital et l’imagerie par résonance magnétique (IRM) ont contribué à l’allongement des délais en vue de la chirurgie.
Conclusion:
Réaliser une biopsie immédiatement après une mammographie anormale, abréger les temps d’attente pour l’IRM et améliorer l’accès en milieu rural pourraient être d’importantes cibles. L’abrègement des temps d’attente en vue d’une chirurgie pour cancer du sein que permet la trajectoire allégée est un argument en faveur d’un système d’aiguillage qui repose sur la coordination de toutes les étapes allant du diagnostic jusqu’au traitement.
Breast cancer is the most common cancer among females, making up 25.6% of all estimated new cancer cases among females in 2023.1 Timely mammograms, biopsies, and subsequent referrals to a surgeon are of key importance in the diagnosis and efficient treatment of breast cancer. However, wait times can be influenced by multiple factors such as patient-level factors, diagnostic pathway, and system-level factors.2–7 Although the issue of treatment delay on breast cancer outcomes has not been entirely resolved, data suggest that pretreatment delays greater than 90 days significantly adversely affect breast cancer mortality.8 Some researchers have suggested that modest pretreatment delays do not adversely affect outcomes;3,9 however, such delays can lead to decreased satisfaction and quality of life, as well as increased anxiety for patients during this considerably stressful period.8,10,11 As such, an effective referral system that ensures patients receive timely coordinated care is crucial.
Studies examining the impact of different models of care on wait times for breast cancer surgery have reported that organized assessment and facilitated care pathways can reduce wait times to breast cancer diagnosis12 and most treatments.13 However, there is a dearth of research examining the impact of different care approaches between care sites serving a mixture of urban and rural patients on breast cancer care outcomes, especially within the Canadian context. The purpose of the present study was to examine 2 distinct referral pathways for female patients diagnosed with breast cancer, identifying potential areas of inefficiencies and other factors that might be addressed to reduce surgical wait times.
Methods
Setting
Distinct models of referral pathways for females who have abnormal mammograms and potential diagnoses of breast cancer exist among care sites operating in one of New Brunswick’s regional health authorities (Figure 1). Mammograms for breast symptoms or cancer screening are read by radiologists in all instances. In the traditional model, mammogram results are then communicated to the patients’ family physician, who discusses the results with the patient and refers them for a biopsy, and then, after receiving the biopsy results, discusses the results with the patient and refers them to a surgeon. The traditional approach to breast cancer was the standard of care offered to patients at Research Site 1, the catchment of a regional hospital covering a large geographic but less populous area that included approximately 50 outlying communities.
Fig. 1.
Comparison of referral pathways for females who have abnormal mammograms and potential diagnoses of breast cancer in the traditional and lean models. Note: FP = family physician.
The lean approach to breast cancer was the standard of care offered to patients living within the catchment of Research Site 2, a tertiary care hospital with a comparatively populous catchment area serving both its own population and that of 20 outlying communities. In the lean model, the radiologist will conduct a biopsy if there are any suspicious findings and will make a referral to a surgeon experienced in treating breast cancer that same day, without waiting for the final biopsy result. The surgeon then coordinates the patient’s care and communicates all decisions to the family physician; the family physician is not responsible for arranging any further investigations or biopsies, making referrals to surgeons or oncologists, or coordinating other care.
Study design
We used a retrospective case review design to analyze the difference in wait times between a lean and traditional approach to patient care. Our primary hypothesis was that the mean time interval of the lean care pathway would be smaller than that of the traditional care pathway. Our secondary hypotheses were that age, distance from research site, size of tumour, breast density, receipt of a screening mammogram, and receipt of magnetic resonance imaging (MRI) would predict overall surgical wait time and wait times at various intervals, and that there would be differences in these factors between care pathways. Tomosynthesis was also expected to reduce wait times but could not be included in the comparison analysis as it was available only at Research Site 1.
Patient population
We examined the electronic files of female patients (aged ≥ 19 yr) diagnosed with stage 0-III breast cancer and surgically treated between February 2016 and July 2019. We excluded patients diagnosed with stage IV disease, those diagnosed but with an unknown or unrecorded stage of disease; those undergoing neoadjuvant therapy; those with a previous history of breast cancer; and those with missing information.
Procedure
Data collected for each patient included date of first abnormal mammogram, date of biopsy, date of first surgical consult, and date of surgery. We calculated time 1 (T1) as the number of days from the first abnormal mammogram to biopsy, time 2 (T2) as the number of days from the biopsy to the first consult with the surgeon, time 3 (T3) as the number of days from the first consult with the surgeon to the surgery, and total time (TT) as the number of days from the first abnormal mammogram to the surgery. Predictors of wait times were also collected from electronic medical records. Predictors included age, mammogram (diagnostic, screening), MRI (yes, no), distance from residence to hospital, size of largest tumour, and breast density. Unfortunately, the approach of describing breast density in terms of categories from A (mostly fatty) to D (extremely dense) was not reported at one of the sites. Therefore, we operationalized breast density as a dichotomous variable (< 50%, > 50%).
Statistical analysis
Variables were summarized using descriptive statistics. We examined histograms, skew, kurtosis, and correlations to ensure that assumptions of the analysis were met.14,15 We used an independent samples t test to test the group differences in wait times. We used a multigroup path analysis (MG-PA) to examine the association of the 6 predictor variables with wait times. A model was specified where T1 was predicted by age, distance from residence to hospital, breast density, and mammogram; T2 was predicted by age, distance from residence to hospital, and MRI; T3 was predicted by age, distance from residence to hospital, MRI, and size of largest tumour; and, finally, TT was regressed on all 6 predictors. We refer to the relationships between variables in the model as parameters and the statistical estimates of the parameters as parameter estimates. We tested both the constrained model (i.e., assuming the parameter estimates are equal for the 2 care pathway approaches) and the unconstrained model (i.e., assuming the parameter estimates are different for the 2 care pathway approaches). We conducted a χ2 difference test between the constrained and unconstrained models, and estimated the parameters for the lean and traditional groups. We assessed model fit — the degree to which the hypothesized model aligned with the data — with a number of fit indices, including the comparative fit index (CFI), the χ2 test, the root mean square error of approximation (RMSEA), and the standardized root mean square residual (SRMR). Tomosynthesis was available only at Research Site 1 and could not be included in the MG-PA; however, because it was a variable of interest, we conducted an exploratory analysis using point-biserial correlations. We used IBM SPSS Statistics 27 for the descriptive analyses and assessing MG-PA assumptions, and Mplus version 8.8 for the MG-PA analysis.
Ethics approval
This study was approved by the Horizon Health Network’s Human Research Protection Program, which is inclusive of the Horizon Health Network Research Ethics Board (RS 2019-2777).
Results
We excluded 418 patients, including 1 with a diagnosis with stage IV disease, 92 with unknown or unrecorded stage of disease, 96 who received neoadjuvant therapy, 118 with a previous history of breast cancer, and 111 with missing information. Of the 456 patients who met the inclusion criteria, patients receiving the lean model of care at Research Site 2 (n = 248) outnumbered patients receiving the traditional model of care at Research Site 1 (n = 208). The age of patients across approaches ranged from 32 to 94 years, with a mean age of 63.7 (standard deviation [SD] 11.1) years. Table 1 provides the results of a descriptive analysis of study variables by care approach.
Table 1.
Descriptive data for patients receiving the traditional and lean models of care
| Characteristic | No. (%) of patients* | ||
|---|---|---|---|
| Overall n = 456 |
Traditional n = 208 |
Lean n = 248 |
|
| Age, yr, mean ± SD | 63.7 ± 11.1 | 63.9 ± 11.2 | 63.6 ± 11.1 |
| Distance from residence to hospital, km, mean ± SD | 36.3 ± 39.0 | 37.5 ± 43.7 | 35.2 ± 34.7 |
| Tumour size, mm, mean ± SD | 19.4 ± 13.2 | 18.2 ± 13.3 | 20.3 ± 13.0 |
| Time, d, mean ± SD | |||
| T1 | 20.0 ± 23.1 | 27.1 ± 25.1 | 14.0 ± 19.4 |
| T2 | 22.5 ± 14.9 | 26.3 ± 16.3 | 19.4 ± 12.8 |
| T3 | 30.9 ± 21.0 | 24.8 ± 14.1 | 36.0 ± 24.3 |
| TT | 73.9 ± 32.5 | 78.9 ± 33.5 | 69.6 ± 31.1 |
| Breast density | |||
| > 50% | 206 (45.2) | 108 (51.9) | 98 (39.5) |
| < 50% | 250 (54.8) | 100 (48.1) | 150 (60.5) |
| Mammogram | |||
| Diagnostic | 214 (46.9) | 94 (45.2) | 120 (48.4) |
| Screening | 242 (53.1) | 114 (54.8) | 128 (51.6) |
| MRI | |||
| No | 280 (61.4) | 99 (47.6) | 181 (73.0) |
| Yes | 176 (38.6) | 109 (52.4) | 67 (27.0) |
MRI = magnetic resonance imaging; SD = standard deviation; T1 = time from abnormal mammogram to biopsy; T2 = time from biopsy to first consult with the surgeon; T3 = time from first consult with surgeon to surgery; TT = total time from abnormal mammogram to surgery.
Unless indicated otherwise.
Our exploratory analysis using point-biserial correlations indicated no statistically significant associations between receiving or not receiving tomosynthesis and T1 (r = −0.09, p = 0.2), T2 (r = −0.07, p = 0.3), T3 (r = 0.02, p = 0.8), or TT (r = −0.10, p = 0.1).
Aside from the CFI (0.980), the constrained model did not fit the data well, according to fit criteria (χ231 = 75.21, p < 0.0001, RMSEA = 0.079, and SRMR = 0.070).16 Conversely, the unconstrained model, which allowed the 2 groups to have different parameter estimates, fit the data well (CFI = 0.999, RMSEA = 0.024, SRMR = 0.026). Further, the χ2 was insignificant (χ214 = 15.78 p = 0.3). Finally, the χ2 difference test between the constrained and unconstrained models indicated that they were significantly different (χ217 = 59.43, p < 0.0001). In sum, the model hypothesizing that there were no differences in wait times between the research sites was not a good fit to the observed data and could not be supported. Conversely, the model hypothesizing differences between the 2 sites was a good fit to the observed data and was significantly different from the model hypothesizing no differences. Therefore, the results support the examination of the parameter estimates in the unconstrained model to determine which predictors were associated with wait times and how these results differed between research sites. Figure 2 and Figure 3 provide the parameter statistics of the unconstrained and constrained models, respectively.
Fig. 2.
Parameter estimates of Research Site 1, which used the traditional model of care. Solid lines indicate significant paths; dashed lines indicate nonsignificant paths. See Related Content for accessible version. Note: MRI = magnetic resonance imaging; T1 = time from abnormal mammogram to biopsy; T2 = time from biopsy to first consult with the surgeon; T3 = time from first consult with surgeon to surgery; TT = total time from abnormal mammogram to surgery.
Fig. 3.
Parameter estimates of Research Site 2, which used the lean model of care. Solid lines indicate significant paths; dashed lines indicate nonsignificant paths. See Related Content for accessible version. Note: T1 = time from abnormal mammogram to biopsy; T2 = time from biopsy to first consult with the surgeon; T3 = time from first consult with surgeon to surgery; TT = total time from abnormal mammogram to surgery.
Time from mammogram to biopsy
There was a significant difference in mean T1 wait times for Research Site 1 (27.1 [SD 25.1] d) and Research Site 2 (14.0 [SD 19.4] d; t386 = 6.17, p < 0.0001). Breast density, age, distance from care centre, and mammogram were examined as predictors of T1. These variables collectively predicted 45% and 21.7% of the variance in T1 for Research Sites 1 and 2, respectively. For both research sites, neither age nor breast density was a statistically significant predictor of T1. Farther distance from residence to research site predicted a longer wait time for the lean care approach at Research Site 2 but not the traditional care approach at Research Site 1 (βtraditional = 0.07, p = 0.2; βlean = 0.24, p < 0.0001). Screening (v. diagnostic) mammograms were associated with longer wait times for both research sites (βtraditional = 1.33, p < 0.0001; βlean = 0.77, p < 0.0001).
Time from biopsy to surgical consult
Patients at Research Site 1 waited a mean of 26.3 (SD 16.3) days at T2, while patients at Research Site 2 waited a mean of 19.4 (SD 12.8) days (t390 = 4.97, p < 0.0001). Age, distance from residence to site, and MRI did not significantly predict the variance in T2 for patients at Research Site 2 (R2 = 0.05, p = 0.07) and predicted only a small proportion of the variance in wait times at T2 (8.5%) for patients at Research Site 1. For patients at Research Site 1, being older (β = 0.16, p = 0.04), living a farther distance from the care site (β = 0.24, p < 0.0001), and receiving an MRI (β = 0.32, p = 0.045) predicted longer T2 wait times.
Time from surgical consult to surgery
Mean T3 wait times were 24.8 (SD 14.1) days for patients at Research Site 1 and 36.0 (SD 24.3) days for patients at Research Site 2 (t407 = 6.16, p < 0.0001). Age, distance from residence to site, size of largest tumour, and MRI collectively predicted 6.5% and 9.3% of the variance in the lean and traditional care approach sites, respectively. Of the 4 predictors, only MRI significantly predicted wait times at T3 (βtraditional = 0.50, p = 0.001; βlean = 0.59, p < 0.0001).
Total time from mammogram to surgery
There was a significant difference in mean wait times for Research Site 1 (78.9 [SD 33.5] d) and Research Site 2 (69.6 [SD 31.1] d) conditions (t454 = 3.07, p = 0.002). Together, the independent variables predicted 31% and 11.3% of the variance in TT wait times at Research Sites 1 and 2, respectively. Age, breast density, and size of largest tumour were not statistically significant predictors of TT for either research site. Screening mammograms (βtraditional = 1.05, p < 0.0001; βlean = 0.48, p < 0.0001), farther distance from residence to hospital (βtraditional = 0.19, p = 0.001; βlean = 0.18, p = 0.003), and receiving an MRI (βtraditional = 0.35, p < 0.0001; βlean = 0.42, p < 0.0001) predicted longer total wait times at both research sites.
Discussion
We compared wait times, as well as key factors associated with wait times, for breast cancer surgery at 2 research sites that used different care pathways (i.e., traditional and lean models of care). Overall, as hypothesized, the traditional model of care had longer surgical wait times than the lean model. For both traditional and lean models of care, receiving a screening mammogram (v. diagnostic mammogram), living a farther distance from the care hospital, and receiving an MRI increased the wait times. In terms of other predictor variables examined, breast density and size of tumour were not associated with wait times in either care model. Age predicted wait times for only Research Site 1 at T2, but the estimate was very low and not likely to be clinically significant.
The finding that distance of residence from hospital significantly predicted time to surgery adds to the literature highlighting differences in patient care depending on whether patients live in rural or urban areas. Greater distance from the treating hospital predicted a longer time to surgery, which may reflect reduced access to health services.17 Patients in outlying communities may receive care that is more consistent with the traditional approach, receiving their screening mammogram in their community health centre first, following up with their family doctor, and then waiting for a referral and biopsy to occur in the urban centre. Within predominantly rural areas, such as New Brunswick, this finding is clinically important to understand to improve care equity. As such, further research is needed to investigate the experiences of patients living at a distance from treatment centres, in rural areas, and in areas where there is a reduced access to primary care.
Supporting previous research, receiving a screening mammogram or MRI also increased wait times from surgical consult to surgery for patients at both sites and from biopsy to surgical consult for the traditional model of care approach only.2,7 In New Brunswick, wait times for screening such as MRIs are sometimes lengthy, so a resultant impact on wait times is to be expected. Likewise, wait times associated with operating room availability are often lengthy and related to hospital-dependent operational procedures, which may be responsible for the increase in wait time for Research Site 2 at T3. Future research should consider the impact of these factors, as well as the effect of staff shortages.
A substantial amount of research has investigated the impact of breast density on imaging for cancer detection and breast cancer risk.18 However, our results indicated that breast density was not a statistically significant predictor of wait time to surgery for either research site, suggesting that timelines were not extended because of breast density. Increased knowledge of the impact of breast density on wait times to surgery would have considerable clinical utility, given the speculation about breast density and potential impacts on breast cancer care. Considering technological advancements in breast cancer screening tools, it is not surprising that breast density did not affect wait times in the current study, but further research is needed to confirm this novel finding.
Regarding imaging, we found that tomosynthesis, which was available only at the research site offering the traditional model of care, was not associated with any of the time variables. Additional research is needed to elucidate this finding.
Although this was a retrospective analysis, a strength of this study is the MG-PA comparison between patients in 2 different facilities with distinct care pathways, increasing the generalizability of the findings and allowing important cross-site comparisons to be made. A mixture of rural and urban patients allowed an examination of the effect of distance from the care site on wait times. The inclusion of predictors and their relation to wait times at multiple time points in the journey from abnormal mammography to surgery is also a strength of this study.
Limitations
The limitations of this study are those that are inherent to a retrospective study and unavailable variables that could influence breast cancer wait times (e.g., race, socioeconomic status). Operating room availability was not considered as a predictor because the data were too variable and difficult to define operationally. Around 20% of the data were missing, with more missing data in the hospital with the traditional model pathway (n = 87) than the hospital with the lean model pathway (n = 24). The missing data may be at least partially owing to differences in the electronic medical record systems between the 2 care sites.
Conclusion
Our finding of shorter wait times to breast cancer surgery in the lean model pathway support the advantage of a direct referral system, wherein organized navigated systems coordinate all aspects of diagnosis and treatment for patients.12,19 We therefore suggest that the traditional model be avoided, given our findings of expedited care and increased efficiency of the lean model. Conducting a biopsy immediately following an abnormal mammogram, improving wait times for MRIs, and improving access for rural patients may be important areas of focus. Continued evaluation and ongoing efforts to reduce wait times and improve process efficiency should be prioritized to optimize patient care (e.g., quality of life11) and breast cancer outcomes.
Footnotes
Competing interests: None declared.
Contributors: Sharon Chiu conceived and designed the work. Natasha Hanson, Leanne Skerry, and Patricia Bryden contributed to data acquisition. Tracy Freeze, Morgan Nesbitt, and Stephen Smith contributed to data analysis and interpretation. Tracy Freeze, Natasha Hanson, and Leanne Skerry drafted the manuscript. All of the authors revised it critically for important intellectual content, gave final approval of the version to be published, and agreed to be accountable for all aspects of the work.
References
- 1.Distribution of cancer cases for selected cancers by sex 2023. Toronto: Dashboard CCS; 2023. Available: https://cancerstats.ca/Incidence/Sex (accessed 2023 May 6). [Google Scholar]
- 2.Bleicher RJ, Ruth K, Sigurdson ER, et al. Preoperative delays in the US Medicare population with breast cancer. J Clin Oncol 2012;30:4485–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Baena-Canada JM, Rodriguez-Perez L, Gamez-Casado S, et al. Evaluation of waiting times for breast cancer diagnosis and surgical treatment. Clin Transl Oncol 2018;20:1345–52. [DOI] [PubMed] [Google Scholar]
- 4.Cordeiro E, Dixon M, Coburn N, et al. A patient-centered approach to wait times in the surgical management of breast cancer in the province of Ontario. Ann Surg Oncol 2015;22:2509–16. [DOI] [PubMed] [Google Scholar]
- 5.Mosunjac M, Park J, Strauss A, et al. Time to treatment for patients receiving BCS in a public and private university hospital in Atlanta. Breast J 2012;18:163–7. [DOI] [PubMed] [Google Scholar]
- 6.Nashed M, Carpenter-Kellett T, Lambert P, et al. Wait time from suspicion to surgery for breast cancer in Manitoba. Cureus 2016;8:e680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Nessim C, Winocour J, Holloway PM, et al. Wait times for breast cancer surgery: effect of magnetic resonance imaging and preoperative investigations on the diagnostic pathway. J Oncol Pract 2015;11:e131–8. [DOI] [PubMed] [Google Scholar]
- 8.Ho PJ, Cook AR, Binte Mohamed Ri NK, et al. Impact of delayed treatment in women diagnosed with breast cancer: a population-based study. Cancer Med 2020;9:2435–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Mansfield SA, Abdel-Rasoul M, Terando AM, et al. Timing of breast cancer surgery-how much does it matter? Breast J 2017;23:444–51. [DOI] [PubMed] [Google Scholar]
- 10.Mathews M, Ryan D, Gadag V, et al. Patient satisfaction with wait-times for breast cancer surgery in Newfoundland and Labrador. Healthc Policy 2016;11:42–53. [PMC free article] [PubMed] [Google Scholar]
- 11.Konieczny M, Cipora E, Roczniak W, et al. Impact of time to initiation of treatment on the quality of life of women with breast cancer. Int J Environ Res Public Health 2020;17:8325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.McKevitt E, Dingee C, Warburton R, et al. Patient navigation reduces time to care for patients with breast symptoms and abnormal screening mammograms. Am J Surg 2018;215:805–11. [DOI] [PubMed] [Google Scholar]
- 13.Blackmore KM, Weerasinghe A, Holloway CMB, et al. Comparison of wait times across the breast cancer treatment pathway among screened women undergoing organized breast assessment versus usual care. Can J Public Health 2019;110:595–605. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Tabachnick BG, Fidell LS. Using Multivariate Statistics, 5th Edition. Boston: Allyn and Bacon; 2007. [Google Scholar]
- 15.Kline RB. Principles and practice of structural equation modeling 2nd edition. New York: Guilford Press; 2005. [Google Scholar]
- 16.Hu Lt, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling 1999;6:1–55. [Google Scholar]
- 17.Febbraro M, Conlon M, Caswell J, et al. Access to cancer care in Northwestern Ontario — a population-based study using administrative data. Current Oncology 2020;27:271–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lian J, Li K. A review of breast density implications and breast cancer screening. Clin Breast Cancer 2020;20:283–90. [DOI] [PubMed] [Google Scholar]
- 19.Chiarelli AM, Muradali D, Blackmore KM, et al. Evaluating wait times from screening to breast cancer diagnosis among women undergoing organised assessment vs usual care. Br J Cancer 2017;116:1254–63. [DOI] [PMC free article] [PubMed] [Google Scholar]