Ann Surg Oncol DOI 10.1245/s10434-015-4626-9
ORIGINAL ARTICLE – BREAST ONCOLOGY
The Value of Ipsilateral Breast Tumor Recurrence as a Quality Indicator: Hospital Variation in the Netherlands M. van der Heiden-van der Loo, PhD1, S. Siesling, PhD1,2, M. W. J. M. Wouters, MD, PhD3, T. van Dalen, MD, PhD4, E. J. T. Rutgers, MD, PhD3, and P. H. M. Peeters, MD, PhD5 Department of Research, Netherlands Comprehensive Cancer Organisation, Utrecht, The Netherlands; 2Department of Health Technology and Services Research, MIRA Institute of Biomedical Science and Technical Medicine, Twente University, Enschede, The Netherlands; 3Department of Surgery, Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands; 4Department of Surgery, Diakonessenhuis Utrecht, Utrecht, The Netherlands; 5 Julius Center of Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands 1
ABSTRACT Purpose. All Dutch hospitals are obliged to report their 5-year ipsilateral breast tumor recurrence (IBTR) rate after breast cancer surgery. Experts decided that these rates should not exceed 5 %. This study determined the value of IBTR as an indicator to compare quality of care between hospitals. Methods. All patients with breast cancer (pT1–3, any N, M0) who underwent surgery in 1 of 92 Dutch hospitals from 2003 to 2006 were identified in the Netherlands Cancer Registry. Data of recurrence was retrieved from hospital records. Five-year IBTR rates for breast-conserving surgery (BCS) and mastectomy were calculated by using the Kaplan–Meier method. Hospital variation was presented in funnel plots. Multivariate analysis was used to assess hospital characteristics associated with IBTR rates. Results. A total of 40,892 breast cancer patients were included. The overall 5-year IBTR rate was 2.85 % (95 % confidence interval 2.68–3.03) and was significantly lower for BCS than for mastectomy (2.38 vs. 3.45 %, p \ 0.001). IBTR rates decreased over time in both groups. Rates varied between 0.77 and 5.70 % between hospitals. When random variation is taken into account, only extremely high IBTR rates can be detected as deviant from the target value of 5 %. Adjusting for tumor and patient characteristics, analyses showed that a higher volume of mastectomies is associated with lower IBTR rates.
Ó Society of Surgical Oncology 2015 First Received: 15 January 2015 M. van der Heiden-van der Loo, PhD e-mail: [email protected]
Conclusions. Our population-based findings show that IBTR rates in the Netherlands are low and have improved over time. The 5-year IBTR rate as an indicator for quality of care of individual hospitals is of limited value.
The call for transparency in the quality of care delivered by individual health care providers has led to continuous attempts to develop valid quality indicators. Quality indicators give insight to specific aspects of care and are used to measure and evaluate care in order to enable improvement. Generally, three types of indicators are recognized: those measuring (1) the organization of care (structural indicators), (2) the care provided (process indicators), and (3) clinical outcome of care (outcome indicators).1–3 To enable health care improvement, indicators need to measure aspects of care that can be influenced by care providers, by improving their skills, altering care processes, or making structural adjustments. Clinical outcome indicators reflect the ‘‘end result’’ of delivered care and therefore often are considered the ultimate instruments to evaluate quality of care.1 However, defining and assessing appropriate clinical outcome indicators has proven to be difficult. If outcome indicators are used to compare clinical care at the hospital level, a clear definition, a uniform collection of the clinical data, and case-mix adjustments are required.3,4 Several indicators have been developed in recent years to evaluate the quality of breast cancer care.5,6 Optimal local disease control with low relapse rates is one of the main objectives of breast cancer treatment. Therefore, ipsilateral breast tumor recurrence (IBTR) rate at 5 years after breast surgery was defined as a quality indicator of the
M. van der Heiden-van der Loo et al.
whole process of breast cancer care. In the Netherlands, based on consensus of experts in the field, the maximum acceptable upper limit value for IBTR within 5 years after breast-conserving surgery (BCS) and mastectomy was set at 5 %.6,7 Since 2009, all Dutch hospitals are legally obliged to report the 5-year IBTR rate to the Dutch Health Care Inspectorate. Detailed clinical data concerning diagnosis, treatment, and survival of breast cancer patients are registered in the Netherlands Cancer Registry (NCR). In a previous study, based on data of patients diagnosed in 2003, we reported that 5-year IBTR rates were 2.6 and 3.5 % for BCS and mastectomy, respectively.8 Currently, 5-year follow-up data of patients treated in 4 consecutive years are available, which enables more in depth analysis of the relevance of this indicator. The purpose of this study was to determine the value of IBTR as an indicator to compare quality of care between hospitals. To meet this objective, we evaluated treatment patterns, time trends, as well as hospital variation of IBTR rates for patients diagnosed with primary operable breast cancer in the Netherlands during the period 2003–2006.
is available through annual linkage with the Municipal Personal Records Database, which has complete information on all deceased and emigrated residents of the Netherlands. The ethics committee of the NCR approved this study. All female breast cancer patients with primary operable breast cancer (T1–3, any N, M0) who underwent BCS or mastectomy in the period January 1, 2003 to December 31, 2006 from all 92 Dutch hospitals were eligible. Patients with previous breast cancer and patients who were not treated with curative intent (patients with macroscopic tumor residue after final surgery; n = 26) were excluded. Patients who underwent more than one operation were classified according to the final surgery. Patients were assigned to the hospital of final surgery. Two hospital characteristics were studied in relation to recurrence rates: volume and type. Hospital volume was based on the total number of patients operated during the period 2003–2006 and divided into quintiles (ranges for BCS: 35–167, 168–229, 230–340, 341–425, 426–677; and mastectomy: 47–128, 129–209, 210–267, 277–340, 341–543). Three types of hospitals were defined: general (n = 54), teaching (n = 29), and academic hospitals (including one specialized oncology center) (n = 9).
PATIENTS AND METHODS Statistical Analysis Study Design and Patients The NCR is a population-based cancer registry, collecting data on all new malignancies in the Netherlands.9 Trained registrars (i.e., data collection staff employed by the NCR) collect data on diagnosis, staging, and treatment from hospital records, including pathology and surgery reports. Stage is coded according to the TNM system of the International Union Against Cancer.10 Follow-up of breast cancer patients is standardized in the Dutch treatment guideline. During the first 5 years after diagnosis, it entails annual follow-up visits to the hospital and includes annual mammography. Registrars were instructed to complete follow-up of all consecutive cases identified at the NCR who were diagnosed from 2003 to 2006 in 1 of 92 Dutch hospitals. Patients without tumor recurrence were coded as tumorfree on the last date of contact in the hospital. If breast cancer reoccurred, date of diagnosis and the localization of recurrence were recorded and categorized as follows: local (in breast, skin or chest wall; both invasive and in situ carcinoma are included), regional (ipsilateral lymph node basin) or distant (metastases), or a combination of these. In case of doubt, registrars were instructed to consult a NCR coding expert or the treating physician. Data on new primary tumors (including new ipsilateral breast tumors) are routinely collected in the NCR. Information on vital status
IBTR included all recurrences irrespective of localization in the breast or histological subtype (i.e., both true recurrences and second primary tumors are included).11 IBTR is noted irrespective of synchronous regional or distant metastases. Although regional and distant metastases are more relevant for the prognosis of the patient, the 5-year IBTR rate is used as an indicator for the quality of local treatment of the breast cancer. Five-year IBTR rates for the period 2003–2006 and for each year separately were calculated by using Kaplan– Meier methods. Patients were censored at date of death (n = 1521), date of diagnosis of a regional or distant metastasis as first event (without local recurrence; n = 403), or date of last visit to the hospital, whichever occurred first. Five-year IBTR rates and 95 % confidence intervals (CI) were calculated for each of the 92 hospitals. If the upper limit of this CI exceeded 5 %, the hospital did not meet the quality limit. This is visualized by setting the upper limit of the indicator in the funnel plot at 5 %. Results of individual hospitals are presented in funnel plots for BCS and mastectomy separately.12 The use of adjuvant treatment (e.g., radiotherapy, chemotherapy, and hormonal treatment) over time was described. Observed trends in rates over time were evaluated by v2 for trend tests.
Hospital Variation in Breast Tumor Recurrences
To further explore the relation between type and volume of the hospital and IBTR rates, multivariate Cox-regression was used, adjusting for patient and tumor characteristics that were related to IBTR rates in a univariate analysis [i.e., year of diagnosis, age, stage, grade, histological subtype, multifocality (i.e., two or more separate tumors), hormone receptor status, HER2 status, tumor residue]. All analyses were conducted using STATA (version 13.1; StataCorp, College Station, TX). RESULTS A total of 40,892 breast cancer patients, diagnosed from 2003 to 2006, were identified. More than 80 % of the tumors were of ductal type and most patients (62 %) had node-negative disease at diagnosis (70 and 50 % in patients who underwent BCS and mastectomy, respectively; Table 1). The number of patients per hospital per year varied from 142 to 322, with an average of 137 patients per year. Overall, 55 % of all patients underwent BCS as final surgery and 45 % had mastectomy. The proportion BCS as final surgery increased from 53 % in 2003 to 56 % in 2006 (p \ 0.001). The percentage of patients with radiotherapy after BCS was high (97 %) and stable over time. After mastectomy, a quarter of patients received radiotherapy. The use of chemotherapy was stable for patients with BCS (33 %) and showed a significant increase only for patients with mastectomy (from 39 to 45 %; p \ 0.001). Hormonal treatment also increased moderately over time in both groups (34–36 % for BCS, p = 0.02 and 46–54 % for mastectomy, p \ 0.001). A total of 1017 patients were diagnosed with an IBTR within 5 years after diagnosis (Table 2). In 213 of these patients (21 %), concurrent regional (n = 59), distant disease (n = 122), or both (n = 32) was diagnosed. The overall 5-year IBTR rate was 2.85 % (95 % CI 2.68–3.03). The IBTR rate was lower for BCS than for mastectomy (2.38 vs. 3.45 %; p \ 0.001). IBTR rate dropped significantly over time (Table 2; p \ 0.001), but the drop was significant only for BCS (p \ 0.001) and not for mastectomy (p = 0.083). In 2006, the 5-year IBTR rate was 1.97 and 2.89 % for patients with BCS and mastectomy, respectively. Results for hospitals are shown in funnel plots for the years 2003–2006 combined (Fig. 1a, b). Overall, IBTR rates for hospitals varied from 0.77 to 5.70 % and stratified for type of surgery from 0 to 5.72 % for BCS and 0 to 12.81 % for mastectomy. Taken random variation into account, for BCS, all hospitals met the preset upper limit of 5 % quality criterion, while for mastectomy only one
hospital showed a higher IBTR rate. This hospital had 75 mastectomies in 4 years; the preset 5 % quality criterion with these numbers was 9.93 % while 12.81 % was observed. Of all patients in this hospital, 75 % had BCS as the final surgery and the IBTR rate for BCS was 1.51 %. Hospital volume (quintiles) or type of hospital (3 categories) did not relate to 5-year IBTR rates for BCS (Table 3). For mastectomies, however, higher hospital volume was associated with lower 5-year IBTR rates: significantly less recurrences were observed in hospitals performing more than 340 mastectomies in 5 years. DISCUSSION This study shows that the quality of local treatment in breast cancer is high and even improved during the years 2003–2006 in the Netherlands. The observed variation in IBTR rates between hospitals is high (0.77–5.70 %), but only one hospital showed only for mastectomies a rate significantly higher than the preset upper limit criterion of 5 %. Five-year IBTR rates for mastectomy were generally slightly higher and showed larger variations between hospitals than for BCS. In addition, only for mastectomies hospital volume seemed important. This suggests that there still is room for improvement, especially for mastectomy. Overall, 5-year IBTR rates in the Netherlands—2.38 % for BCS and 3.45 % for mastectomy—were significantly lower than the ‘‘experts’’ predefined value of 5 %. Rates also were lower than those reported in large meta-analyses. The EBCTCG overview reports on data from over 42,000 patients in 78 trials, started in 1995. Five-year isolated breast recurrence rates were 2.3 % (with radiotherapy) and 6.3 % (without radiotherapy) for mastectomies with nodenegative disease, whereas these values were 5.8 and 22.8 % for mastectomy in node-positive disease.13 In our study, the 5-year IBTR rate for mastectomy was 3.45 with 50 % node-negative disease and 27 % had radiotherapy. For BCS, the EBCTCG overview reported 5-year IBTR rates of 6.7 % for node-negative patients with radiotherapy and 11.0 % in patients with node-positive disease.13 IBTR rate was 2.38 % in our study; 97 % of patients received radiotherapy. The authors of the EBCTCG overview predicted lower recurrence rates for the future due to advancements in screening and treatment.14 Hospitals varied in overall IBTR rates from 0.77 to 5.70 %. Hospital variation was considerably higher for mastectomy than for BCS (0–12.81 and 0–5.72 %, respectively). The upper limit for acceptable IBTR (5 %, 5 years after diagnosis) as defined by a working committee in 2006 and derived from EUSOMA guidelines published in 2001 was achieved when combining data from all hospitals.7,13 Although the observed percentages of 5-year
M. van der Heiden-van der Loo et al. TABLE 1 Patient and tumor characteristics according to final surgery in 40,892 breast cancer patients, diagnosed 2003–2006 in 92 hospitals in the Netherlands BCS (n = 22,450) N
Mastectomy (n = 18,442) %
N
%
Age (years) \35 35–49 50–69 C70
372
2
511
3
5335 12,924
24 57
4478 7719
24 42
3819
17
5734
31 4
a
pT (cm) B0.5 (BT1A)
906
4
724
[0.5–1 (T1B)
4282
19
1443
8
11,244
50
5833
32
2–5 (T2)
5854
26
8814
48
[5 (T3)
120
1
1483
8
1–2 (T1C)
Unknown
44
145
pN 0
15,356
70
8622
50
1
5133
24
5494
32
2
889
4
1953
11
3
360
2
1133
7
Unknown Grade
712
1240
Well differentiated
5524
27
2663
16
Moderately differentiated
9254
44
7652
46
Poorly differentiated
6129
29
6355
38
Unknown
1543
1772
Histological subtype Ductal Lobular Ductal ? lobular b
Other
18,636
83
14,238
77
1854
8
2561
14
764
4
919
5
1196
5
724
4
Multifocality No
16,589
94
10,980
75
Yes
1139
6
3748
25
Unknownc
4722
3714
Expression of receptors ER or PR positive ER and PR negative Unknownc
15,878 3188
83 17
3384
12,504 3184
80 20
2754
Her2 No
9884
88
7405
82
Yes
1408
12
1624
18
Unknownd
11,158
9413
Residue No Microscopic Unknown
c
20,850
95
17,280
1043
5
563
557
97 3
599
Hospital volume 2003–2006e Quintile 1
3975
18
2926
16
Hospital Variation in Breast Tumor Recurrences TABLE 1 continued BCS (n = 22,450) N
Mastectomy (n = 18,442) %
N
%
Quintile 2
4519
20
3884
21
Quintile 3
4255
19
4213
23
Quintile 4
4930
22
3639
20
Quintile 5
4771
21
3780
20
Hospital type General Teaching Academic a
9857
44
9129
49
10,577
47
7708
42
2016
9
1605
9
Clinical T-stage if pathological T-stage was missing or patient underwent neoadjuvant therapy
b
Histological subtype ‘‘other’’: e.g., mucinous adenocarcinoma, medullary carcinoma, metaplastic carcinoma
c
Category unknown is largely due to missing in registration in earlier years
d
Her2 testing became standard in September 2005. Category unknown consists largely of missing data in earlier years
e
Quintiles based on the total number of specified surgeries per hospital during the period 2003–2006. BCS: 35–167, 168–229, 230–340, 341–425, 426–677; mastectomy: 47–128, 129–209, 210–267, 277–340, 341–543
TABLE 2 Five-year IBTR rates according to final surgery and number of hospitals with rates [5 %, observed and preset criterion PYa
N
IBTR n
IBTR (95 % CI)b
Hospitals observed [5 % N (%)
Hospitals preset 5 % N (%)c
Breast-conserving surgery 2003
5329
24,240
135
2.80 % (2.37–3.31)
19 (21)
1 (1)
2004
5520
24,688
136
2.84 % (2.40–3.36)
14 (15)
2 (2)
2005
5694
25,929
101
1.98 % (1.63–2.41)
8 (9)
0 (0)
2006
5907
26,928
105
1.97 % (1.63–2.38)
8 (9)
0 (0)
22,450
101,785
477
2.38 % (2.18–2.60)
2 (2)
0 (0)
2003
4641
19,134
142
3.59 % (3.05–4.22)
28 (30)
5 (5)
2004
4690
18,929
151
3.91 % (3.33–4.58)
30 (33)
1 (1)
2005
4461
18,448
131
3.44 % (2.90–4.08)
29 (32)
0 (0)
2006
4650
19,431
116
2.89 % (2.41–3.46)
20 (22)
3 (3)
18,442
75,942
540
3.45 % (3.18–3.75)
19 (21)
1 (1)
2003–2006 Mastectomy
2003–2006 a
Person years of follow-up
b
Kaplan–Meier estimate
c
Based on 95 % CI around the target value of 5 %
IBTR rates for several hospitals exceeded the predefined 5 % boarder, for the correct interpretation of these values the random variation has to be taken into account. The result is that with lower number of surgeries in a hospital the preset 5 % upper limit criterion will increase. This can be seen clearly in the funnel plots where the upper limit criterion is included as the upper line. The consequence is that with low volumes the observed IBTR rates can vary hugely without reaching the preset criterion. For example, hospitals with 100 procedures may have up to 10 % recurrences in 5 years and still meet the preset criterion of 5 % of recurrences in 5 years; with only 50 procedures
even recurrence rates of 15 % in 5 years are within acceptable ranges. So, the consequence is that only extreme high recurrence rates can be detected as deviant. This is an important shortcoming of this indicator. In addition, information to compute this indicator is available only after many years so any adjustments to improve are too late. Furthermore, in addition to random variation there may be incomparability of the hospital-specific rates due to confounding by patient and tumor characteristics. Case mix correction is of importance when interpreting hospital variation in outcome indicators.12,15 However, with the small numbers of patients and a low IBTR rate, case mix
M. van der Heiden-van der Loo et al.
a
20%
15%
IBTR rate
FIG. 1 a BCS: funnel plot with ipsilateral breast tumor recurrence rates for BCS in 2003–2006 by hospital. b Mastectomy: funnel plot with Ipsilateral breast tumor recurrence rates for mastectomy in 2003–2006 by hospital
10%
5%
0% 0
100
200
300
400
500
600
700
600
700
Number of patients with BCS as final surgery per hospital 95% Cl
b
BCS
20%
IBTR rate
15%
10%
5%
0% 0
100
200
400
300
500
Number of patients with mastectomy as final surgery per hospital 95% Cl
Mastectomy
TABLE 3 Multivariate analyses of hospital characteristics associated with ipsilateral breast tumor recurrence in the Netherlands BCS, adjusted model Aa
BCS, adjusted model B
Mastectomy, adjusted model A
Mastectomy, adjusted model B
HR
HR
95 % CI
HR
HR
95 % CI
95 % CI
95 % CI
b
Hospital volume 2003–2006 Quintile 1
1 (ref)
Quintile 2
1.14
0.83–1.55
1.11
1 (ref) 0.81–1.51
0.79
1 (ref) 0.61–1.04
1 (ref) 0.77
0.59–1.01
Quintile 3
1.32
0.97–1.80
1.32
0.97–1.80
0.87
0.67–1.13
0.84
0.65–1.10
Quintile 4
1.21
0.85–1.74
1.28
0.90–1.84
0.70
0.52–0.94#
0.69
0.51–0.92#
Quintile 5
1.17
0.81–1.69
1.29
0.89–1.87
0.76
0.57–1.02
0.73
0.55–0.98#
Hospital type General
1 (ref)
Teaching
1.03
0.80–1.34
0.98
1 (ref) 0.76–1.28
0.94
1 (ref) 0.77–1.15
1 (ref) 0.94
0.77–1.14
Academic
0.92
0.65–1.31
0.86
0.60–1.23
0.93
0.68–1.27
0.93
0.68–1.28
a
Model A included adjustment for the following variables (as described in Table 1 unless otherwise specified): year of diagnosis, age, T (T2 and T3 combined), N (N2 and N3 combined), grade, histological subtype, multifocality, receptor status, HER2 status and residue; model B included adjustment model A and radiation therapy, chemotherapy and hormonal treatment b
Divided into quintiles based on the total number of specified surgeries per hospital over the period 2003–2006. BCS: 35–167, 168–229, 230–340, 341–425, 426–677; mastectomy: 47–128, 129–209, 210–267, 277–340, 341–543
#
Significant
correction is problematic, especially given the multiple predictors needed. We conclude that this is not a valid indicator to judge the overall quality of breast cancer care of an individual hospital. Another consideration is whether a follow-up of 5 years is long enough to detect all IBTR. Approximately threequarters of all IBTR are diagnosed within 5 years after initial diagnosis.16 Late recurrences (after 5 years) are
more often new, primary tumors, which are associated with better prognosis.17,18 However, the indicator aims to measure the optimal local control of the diagnosed cancer, so true recurrences (mostly occurring within 5 years) are of more interest than second primary breast cancers. The major strength of our study is the availability of a large population based cohort of all breast cancer patients diagnosed in 4 consecutive years in all 92 Dutch hospitals.
Hospital Variation in Breast Tumor Recurrences
The data used in this study was collected by trained registrars and based on a strict coding manual. In case of doubt, registrars were instructed to consult with a registration expert, which minimizes misclassification. In concluding that the IBTR rate is not a suitable outcome indicator for comparing the quality of care between hospitals, the question arises of which indicators do? Process indicators could be of help. A focus on the violation of agreed standards in specific patients may minimize the need for case mix adjustment. These indicators can be quickly available and also pinpoint the target for quick improvements.5,19 The nationwide clinical breast cancer audit in the Netherlands comprises 30 quality indicators, including process indicators (e.g., measuring the proportion of new diagnoses that are discussed in a preoperative multidisciplinary team meeting and the proportion of incomplete first surgeries). The first promising results of this audit were recently presented at the 37th Annual San Antonio Breast Cancer Symposium.20,21 Our findings show that IBTR rates are low in the Netherlands and have substantially improved in recent years. This outcome indicator is of interest on a national level for international comparison but is not suitable for comparing hospital quality of care, because the number of treated patients per hospital is limited and the number of recurrences is too small in individual hospitals. CONFLICT OF INTEREST no conflict of interest.
9.
10.
11.
12. 13.
14.
15.
16.
The authors declare that they have 17.
REFERENCES 1. Birkmeyer JD, Dimick JB, Birkmeyer NJ. Measuring the quality of surgical care: structure, process, or outcomes? J Am Coll Surg. 2004;198(4):626–32. 2. Donabedian A: Evaluating the quality of medical care. Milbank Q. 1966;83(4):691–729. 3. Rubin HR, Pronovost P, Diette GB. From a process of care to a measure: the development and testing of a quality indicator. Int J Qual Health Care. 2001;13(6):489–96. 4. Brien SE, Dixon E, Ghali WA. Measuring and reporting on quality in health care: a framework and road map for improving care. J Surg Oncol. 2009;99(8):462–6. 5. Del Turco MR, Ponti A, Bick U, Biganzoli L, Cserni G, Cutuli B, Decker T, Dietel M, Gentilini O, Kuehn T, et al. Quality indicators in breast cancer care. Eur J Cancer. 2010;46(13):2344–56. 6. Rutgers EJ, Wittenberg J, Kuijpers AC. Kwaliteit van zorg rond mammacarcinoom in de etalage. In: Utrecht: CBO, Orde van Medisch Specialisten; 2006. 7. Rutgers EJ, Group EC. Quality control in the locoregional treatment of breast cancer. Eur J Cancer. 2001;37(4):447–53. 8. van der Heiden-van der Loo M, Ho VK, Damhuis RA, Siesling S, Menke MB, Peeters PH, Rutgers EJ: Percentage of local recurrence following treatment for breast cancer is not a suitable
18.
19.
20.
21.
performance indicator. Nederlands Tijdschrift Geneeskunde. 2010;154:A1984. Schouten LJ, Hoppener P, van den Brandt PA, Knottnerus JA, Jager JJ. Completeness of cancer registration in Limburg, The Netherlands. Int J Epidemiol. 1993;22(3):369–76. Sobin LH, Wittekind C. TNM Classification of malignant tumours. International Union against cancer (UICC), vol. 6. New York: Wiley-Liss; 2002. Moossdorff M, van Roozendaal LM, Strobbe LJ, Aebi S, Cameron DA, Dixon JM, Giuliano AE, Haffty BG, Hickey BE, Hudis CA, et al. Maastricht delphi consensus on event definitions for classification of recurrence in breast cancer research. J Natl Cancer Inst. 2014;106(12):dju288. Spiegelhalter DJ. Funnel plots for comparing institutional performance. Stat Med. 2005;24(8):1185–202. Clarke M, Collins R, Darby S, Davies C, Elphinstone P, Evans E, Godwin J, Gray R, Hicks C, James S, et al. Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet. 2005;366(9503):2087–106. Early Breast Cancer Trialists’ Collaborative G, Darby S, McGale P, Correa C, Taylor C, Arriagada R, Clarke M, Cutter D, Davies C, Ewertz M, et al. Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet. 2011;378(9804):1707–16. Kolfschoten NE, Marang van de Mheen PJ, Gooiker GA, Eddes EH, Kievit J, Tollenaar RA, Wouters MW, Dutch Surgical Colorectal Audit Group. Variation in case-mix between hospitals treating colorectal cancer patients in the Netherlands. Eur J Surg Oncol. 2011;37(11):956–63. Park CC, Mitsumori M, Nixon A, Recht A, Connolly J, Gelman R, Silver B, Hetelekidis S, Abner A, Harris JR, et al. Outcome at 8 years after breast-conserving surgery and radiation therapy for invasive breast cancer: influence of margin status and systemic therapy on local recurrence. J Clin Oncol. 2000;18(8):1668–75. Moran MS, Haffty BG. Local-regional breast cancer recurrence: prognostic groups based on patterns of failure. Breast J. 2002;8(2):81–7. Panet-Raymond V, Truong PT, McDonald RE, Alexander C, Ross L, Ryhorchuk A, Watson PH. True recurrence versus new primary: an analysis of ipsilateral breast tumor recurrences after breast-conserving therapy. Int J Radiat Oncol Biol Phys. 2011;81(2):409–17. Lilford R, Mohammed MA, Spiegelhalter D, Thomson R. Use and misuse of process and outcome data in managing performance of acute medical care: avoiding institutional stigma. Lancet. 2004;363(9415):1147–54. Tjan-Heijnen VC, Van Bommel ACM, van der Heiden-van der Loo M, Westenend PJ, De Vries B, Smorenburg CH, Jager A, Lobbes MB, Pijnappel RM, Maduro JH, et al. Multidisciplinary breast cancer care registry and quality control system in the Netherlands: The NABON breast cancer audit. In 37th Annual San Antonio Breast Cancer Symposium. San Antonio;2014. Van Bommel ACM, Vrancken Peeters MJ, van der Heiden-van der Loo M, Van Dalen T, Rutgers E, Wouters MW, Lobbes MB, Pijnappel RM, Mureau MA, Westenend PJ, et al. The NABON breast cancer audit; quality improvement in three years’ time. In 27th Annual San Antonio Breast Cancer Symposium. San Antonio;2014.
ORIGINAL ARTICLE – BREAST ONCOLOGY
The Value of Ipsilateral Breast Tumor Recurrence as a Quality Indicator: Hospital Variation in the Netherlands M. van der Heiden-van der Loo, PhD1, S. Siesling, PhD1,2, M. W. J. M. Wouters, MD, PhD3, T. van Dalen, MD, PhD4, E. J. T. Rutgers, MD, PhD3, and P. H. M. Peeters, MD, PhD5 Department of Research, Netherlands Comprehensive Cancer Organisation, Utrecht, The Netherlands; 2Department of Health Technology and Services Research, MIRA Institute of Biomedical Science and Technical Medicine, Twente University, Enschede, The Netherlands; 3Department of Surgery, Netherlands Cancer Institute-Antoni van Leeuwenhoek Hospital, Amsterdam, The Netherlands; 4Department of Surgery, Diakonessenhuis Utrecht, Utrecht, The Netherlands; 5 Julius Center of Health Sciences and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands 1
ABSTRACT Purpose. All Dutch hospitals are obliged to report their 5-year ipsilateral breast tumor recurrence (IBTR) rate after breast cancer surgery. Experts decided that these rates should not exceed 5 %. This study determined the value of IBTR as an indicator to compare quality of care between hospitals. Methods. All patients with breast cancer (pT1–3, any N, M0) who underwent surgery in 1 of 92 Dutch hospitals from 2003 to 2006 were identified in the Netherlands Cancer Registry. Data of recurrence was retrieved from hospital records. Five-year IBTR rates for breast-conserving surgery (BCS) and mastectomy were calculated by using the Kaplan–Meier method. Hospital variation was presented in funnel plots. Multivariate analysis was used to assess hospital characteristics associated with IBTR rates. Results. A total of 40,892 breast cancer patients were included. The overall 5-year IBTR rate was 2.85 % (95 % confidence interval 2.68–3.03) and was significantly lower for BCS than for mastectomy (2.38 vs. 3.45 %, p \ 0.001). IBTR rates decreased over time in both groups. Rates varied between 0.77 and 5.70 % between hospitals. When random variation is taken into account, only extremely high IBTR rates can be detected as deviant from the target value of 5 %. Adjusting for tumor and patient characteristics, analyses showed that a higher volume of mastectomies is associated with lower IBTR rates.
Ó Society of Surgical Oncology 2015 First Received: 15 January 2015 M. van der Heiden-van der Loo, PhD e-mail: [email protected]
Conclusions. Our population-based findings show that IBTR rates in the Netherlands are low and have improved over time. The 5-year IBTR rate as an indicator for quality of care of individual hospitals is of limited value.
The call for transparency in the quality of care delivered by individual health care providers has led to continuous attempts to develop valid quality indicators. Quality indicators give insight to specific aspects of care and are used to measure and evaluate care in order to enable improvement. Generally, three types of indicators are recognized: those measuring (1) the organization of care (structural indicators), (2) the care provided (process indicators), and (3) clinical outcome of care (outcome indicators).1–3 To enable health care improvement, indicators need to measure aspects of care that can be influenced by care providers, by improving their skills, altering care processes, or making structural adjustments. Clinical outcome indicators reflect the ‘‘end result’’ of delivered care and therefore often are considered the ultimate instruments to evaluate quality of care.1 However, defining and assessing appropriate clinical outcome indicators has proven to be difficult. If outcome indicators are used to compare clinical care at the hospital level, a clear definition, a uniform collection of the clinical data, and case-mix adjustments are required.3,4 Several indicators have been developed in recent years to evaluate the quality of breast cancer care.5,6 Optimal local disease control with low relapse rates is one of the main objectives of breast cancer treatment. Therefore, ipsilateral breast tumor recurrence (IBTR) rate at 5 years after breast surgery was defined as a quality indicator of the
M. van der Heiden-van der Loo et al.
whole process of breast cancer care. In the Netherlands, based on consensus of experts in the field, the maximum acceptable upper limit value for IBTR within 5 years after breast-conserving surgery (BCS) and mastectomy was set at 5 %.6,7 Since 2009, all Dutch hospitals are legally obliged to report the 5-year IBTR rate to the Dutch Health Care Inspectorate. Detailed clinical data concerning diagnosis, treatment, and survival of breast cancer patients are registered in the Netherlands Cancer Registry (NCR). In a previous study, based on data of patients diagnosed in 2003, we reported that 5-year IBTR rates were 2.6 and 3.5 % for BCS and mastectomy, respectively.8 Currently, 5-year follow-up data of patients treated in 4 consecutive years are available, which enables more in depth analysis of the relevance of this indicator. The purpose of this study was to determine the value of IBTR as an indicator to compare quality of care between hospitals. To meet this objective, we evaluated treatment patterns, time trends, as well as hospital variation of IBTR rates for patients diagnosed with primary operable breast cancer in the Netherlands during the period 2003–2006.
is available through annual linkage with the Municipal Personal Records Database, which has complete information on all deceased and emigrated residents of the Netherlands. The ethics committee of the NCR approved this study. All female breast cancer patients with primary operable breast cancer (T1–3, any N, M0) who underwent BCS or mastectomy in the period January 1, 2003 to December 31, 2006 from all 92 Dutch hospitals were eligible. Patients with previous breast cancer and patients who were not treated with curative intent (patients with macroscopic tumor residue after final surgery; n = 26) were excluded. Patients who underwent more than one operation were classified according to the final surgery. Patients were assigned to the hospital of final surgery. Two hospital characteristics were studied in relation to recurrence rates: volume and type. Hospital volume was based on the total number of patients operated during the period 2003–2006 and divided into quintiles (ranges for BCS: 35–167, 168–229, 230–340, 341–425, 426–677; and mastectomy: 47–128, 129–209, 210–267, 277–340, 341–543). Three types of hospitals were defined: general (n = 54), teaching (n = 29), and academic hospitals (including one specialized oncology center) (n = 9).
PATIENTS AND METHODS Statistical Analysis Study Design and Patients The NCR is a population-based cancer registry, collecting data on all new malignancies in the Netherlands.9 Trained registrars (i.e., data collection staff employed by the NCR) collect data on diagnosis, staging, and treatment from hospital records, including pathology and surgery reports. Stage is coded according to the TNM system of the International Union Against Cancer.10 Follow-up of breast cancer patients is standardized in the Dutch treatment guideline. During the first 5 years after diagnosis, it entails annual follow-up visits to the hospital and includes annual mammography. Registrars were instructed to complete follow-up of all consecutive cases identified at the NCR who were diagnosed from 2003 to 2006 in 1 of 92 Dutch hospitals. Patients without tumor recurrence were coded as tumorfree on the last date of contact in the hospital. If breast cancer reoccurred, date of diagnosis and the localization of recurrence were recorded and categorized as follows: local (in breast, skin or chest wall; both invasive and in situ carcinoma are included), regional (ipsilateral lymph node basin) or distant (metastases), or a combination of these. In case of doubt, registrars were instructed to consult a NCR coding expert or the treating physician. Data on new primary tumors (including new ipsilateral breast tumors) are routinely collected in the NCR. Information on vital status
IBTR included all recurrences irrespective of localization in the breast or histological subtype (i.e., both true recurrences and second primary tumors are included).11 IBTR is noted irrespective of synchronous regional or distant metastases. Although regional and distant metastases are more relevant for the prognosis of the patient, the 5-year IBTR rate is used as an indicator for the quality of local treatment of the breast cancer. Five-year IBTR rates for the period 2003–2006 and for each year separately were calculated by using Kaplan– Meier methods. Patients were censored at date of death (n = 1521), date of diagnosis of a regional or distant metastasis as first event (without local recurrence; n = 403), or date of last visit to the hospital, whichever occurred first. Five-year IBTR rates and 95 % confidence intervals (CI) were calculated for each of the 92 hospitals. If the upper limit of this CI exceeded 5 %, the hospital did not meet the quality limit. This is visualized by setting the upper limit of the indicator in the funnel plot at 5 %. Results of individual hospitals are presented in funnel plots for BCS and mastectomy separately.12 The use of adjuvant treatment (e.g., radiotherapy, chemotherapy, and hormonal treatment) over time was described. Observed trends in rates over time were evaluated by v2 for trend tests.
Hospital Variation in Breast Tumor Recurrences
To further explore the relation between type and volume of the hospital and IBTR rates, multivariate Cox-regression was used, adjusting for patient and tumor characteristics that were related to IBTR rates in a univariate analysis [i.e., year of diagnosis, age, stage, grade, histological subtype, multifocality (i.e., two or more separate tumors), hormone receptor status, HER2 status, tumor residue]. All analyses were conducted using STATA (version 13.1; StataCorp, College Station, TX). RESULTS A total of 40,892 breast cancer patients, diagnosed from 2003 to 2006, were identified. More than 80 % of the tumors were of ductal type and most patients (62 %) had node-negative disease at diagnosis (70 and 50 % in patients who underwent BCS and mastectomy, respectively; Table 1). The number of patients per hospital per year varied from 142 to 322, with an average of 137 patients per year. Overall, 55 % of all patients underwent BCS as final surgery and 45 % had mastectomy. The proportion BCS as final surgery increased from 53 % in 2003 to 56 % in 2006 (p \ 0.001). The percentage of patients with radiotherapy after BCS was high (97 %) and stable over time. After mastectomy, a quarter of patients received radiotherapy. The use of chemotherapy was stable for patients with BCS (33 %) and showed a significant increase only for patients with mastectomy (from 39 to 45 %; p \ 0.001). Hormonal treatment also increased moderately over time in both groups (34–36 % for BCS, p = 0.02 and 46–54 % for mastectomy, p \ 0.001). A total of 1017 patients were diagnosed with an IBTR within 5 years after diagnosis (Table 2). In 213 of these patients (21 %), concurrent regional (n = 59), distant disease (n = 122), or both (n = 32) was diagnosed. The overall 5-year IBTR rate was 2.85 % (95 % CI 2.68–3.03). The IBTR rate was lower for BCS than for mastectomy (2.38 vs. 3.45 %; p \ 0.001). IBTR rate dropped significantly over time (Table 2; p \ 0.001), but the drop was significant only for BCS (p \ 0.001) and not for mastectomy (p = 0.083). In 2006, the 5-year IBTR rate was 1.97 and 2.89 % for patients with BCS and mastectomy, respectively. Results for hospitals are shown in funnel plots for the years 2003–2006 combined (Fig. 1a, b). Overall, IBTR rates for hospitals varied from 0.77 to 5.70 % and stratified for type of surgery from 0 to 5.72 % for BCS and 0 to 12.81 % for mastectomy. Taken random variation into account, for BCS, all hospitals met the preset upper limit of 5 % quality criterion, while for mastectomy only one
hospital showed a higher IBTR rate. This hospital had 75 mastectomies in 4 years; the preset 5 % quality criterion with these numbers was 9.93 % while 12.81 % was observed. Of all patients in this hospital, 75 % had BCS as the final surgery and the IBTR rate for BCS was 1.51 %. Hospital volume (quintiles) or type of hospital (3 categories) did not relate to 5-year IBTR rates for BCS (Table 3). For mastectomies, however, higher hospital volume was associated with lower 5-year IBTR rates: significantly less recurrences were observed in hospitals performing more than 340 mastectomies in 5 years. DISCUSSION This study shows that the quality of local treatment in breast cancer is high and even improved during the years 2003–2006 in the Netherlands. The observed variation in IBTR rates between hospitals is high (0.77–5.70 %), but only one hospital showed only for mastectomies a rate significantly higher than the preset upper limit criterion of 5 %. Five-year IBTR rates for mastectomy were generally slightly higher and showed larger variations between hospitals than for BCS. In addition, only for mastectomies hospital volume seemed important. This suggests that there still is room for improvement, especially for mastectomy. Overall, 5-year IBTR rates in the Netherlands—2.38 % for BCS and 3.45 % for mastectomy—were significantly lower than the ‘‘experts’’ predefined value of 5 %. Rates also were lower than those reported in large meta-analyses. The EBCTCG overview reports on data from over 42,000 patients in 78 trials, started in 1995. Five-year isolated breast recurrence rates were 2.3 % (with radiotherapy) and 6.3 % (without radiotherapy) for mastectomies with nodenegative disease, whereas these values were 5.8 and 22.8 % for mastectomy in node-positive disease.13 In our study, the 5-year IBTR rate for mastectomy was 3.45 with 50 % node-negative disease and 27 % had radiotherapy. For BCS, the EBCTCG overview reported 5-year IBTR rates of 6.7 % for node-negative patients with radiotherapy and 11.0 % in patients with node-positive disease.13 IBTR rate was 2.38 % in our study; 97 % of patients received radiotherapy. The authors of the EBCTCG overview predicted lower recurrence rates for the future due to advancements in screening and treatment.14 Hospitals varied in overall IBTR rates from 0.77 to 5.70 %. Hospital variation was considerably higher for mastectomy than for BCS (0–12.81 and 0–5.72 %, respectively). The upper limit for acceptable IBTR (5 %, 5 years after diagnosis) as defined by a working committee in 2006 and derived from EUSOMA guidelines published in 2001 was achieved when combining data from all hospitals.7,13 Although the observed percentages of 5-year
M. van der Heiden-van der Loo et al. TABLE 1 Patient and tumor characteristics according to final surgery in 40,892 breast cancer patients, diagnosed 2003–2006 in 92 hospitals in the Netherlands BCS (n = 22,450) N
Mastectomy (n = 18,442) %
N
%
Age (years) \35 35–49 50–69 C70
372
2
511
3
5335 12,924
24 57
4478 7719
24 42
3819
17
5734
31 4
a
pT (cm) B0.5 (BT1A)
906
4
724
[0.5–1 (T1B)
4282
19
1443
8
11,244
50
5833
32
2–5 (T2)
5854
26
8814
48
[5 (T3)
120
1
1483
8
1–2 (T1C)
Unknown
44
145
pN 0
15,356
70
8622
50
1
5133
24
5494
32
2
889
4
1953
11
3
360
2
1133
7
Unknown Grade
712
1240
Well differentiated
5524
27
2663
16
Moderately differentiated
9254
44
7652
46
Poorly differentiated
6129
29
6355
38
Unknown
1543
1772
Histological subtype Ductal Lobular Ductal ? lobular b
Other
18,636
83
14,238
77
1854
8
2561
14
764
4
919
5
1196
5
724
4
Multifocality No
16,589
94
10,980
75
Yes
1139
6
3748
25
Unknownc
4722
3714
Expression of receptors ER or PR positive ER and PR negative Unknownc
15,878 3188
83 17
3384
12,504 3184
80 20
2754
Her2 No
9884
88
7405
82
Yes
1408
12
1624
18
Unknownd
11,158
9413
Residue No Microscopic Unknown
c
20,850
95
17,280
1043
5
563
557
97 3
599
Hospital volume 2003–2006e Quintile 1
3975
18
2926
16
Hospital Variation in Breast Tumor Recurrences TABLE 1 continued BCS (n = 22,450) N
Mastectomy (n = 18,442) %
N
%
Quintile 2
4519
20
3884
21
Quintile 3
4255
19
4213
23
Quintile 4
4930
22
3639
20
Quintile 5
4771
21
3780
20
Hospital type General Teaching Academic a
9857
44
9129
49
10,577
47
7708
42
2016
9
1605
9
Clinical T-stage if pathological T-stage was missing or patient underwent neoadjuvant therapy
b
Histological subtype ‘‘other’’: e.g., mucinous adenocarcinoma, medullary carcinoma, metaplastic carcinoma
c
Category unknown is largely due to missing in registration in earlier years
d
Her2 testing became standard in September 2005. Category unknown consists largely of missing data in earlier years
e
Quintiles based on the total number of specified surgeries per hospital during the period 2003–2006. BCS: 35–167, 168–229, 230–340, 341–425, 426–677; mastectomy: 47–128, 129–209, 210–267, 277–340, 341–543
TABLE 2 Five-year IBTR rates according to final surgery and number of hospitals with rates [5 %, observed and preset criterion PYa
N
IBTR n
IBTR (95 % CI)b
Hospitals observed [5 % N (%)
Hospitals preset 5 % N (%)c
Breast-conserving surgery 2003
5329
24,240
135
2.80 % (2.37–3.31)
19 (21)
1 (1)
2004
5520
24,688
136
2.84 % (2.40–3.36)
14 (15)
2 (2)
2005
5694
25,929
101
1.98 % (1.63–2.41)
8 (9)
0 (0)
2006
5907
26,928
105
1.97 % (1.63–2.38)
8 (9)
0 (0)
22,450
101,785
477
2.38 % (2.18–2.60)
2 (2)
0 (0)
2003
4641
19,134
142
3.59 % (3.05–4.22)
28 (30)
5 (5)
2004
4690
18,929
151
3.91 % (3.33–4.58)
30 (33)
1 (1)
2005
4461
18,448
131
3.44 % (2.90–4.08)
29 (32)
0 (0)
2006
4650
19,431
116
2.89 % (2.41–3.46)
20 (22)
3 (3)
18,442
75,942
540
3.45 % (3.18–3.75)
19 (21)
1 (1)
2003–2006 Mastectomy
2003–2006 a
Person years of follow-up
b
Kaplan–Meier estimate
c
Based on 95 % CI around the target value of 5 %
IBTR rates for several hospitals exceeded the predefined 5 % boarder, for the correct interpretation of these values the random variation has to be taken into account. The result is that with lower number of surgeries in a hospital the preset 5 % upper limit criterion will increase. This can be seen clearly in the funnel plots where the upper limit criterion is included as the upper line. The consequence is that with low volumes the observed IBTR rates can vary hugely without reaching the preset criterion. For example, hospitals with 100 procedures may have up to 10 % recurrences in 5 years and still meet the preset criterion of 5 % of recurrences in 5 years; with only 50 procedures
even recurrence rates of 15 % in 5 years are within acceptable ranges. So, the consequence is that only extreme high recurrence rates can be detected as deviant. This is an important shortcoming of this indicator. In addition, information to compute this indicator is available only after many years so any adjustments to improve are too late. Furthermore, in addition to random variation there may be incomparability of the hospital-specific rates due to confounding by patient and tumor characteristics. Case mix correction is of importance when interpreting hospital variation in outcome indicators.12,15 However, with the small numbers of patients and a low IBTR rate, case mix
M. van der Heiden-van der Loo et al.
a
20%
15%
IBTR rate
FIG. 1 a BCS: funnel plot with ipsilateral breast tumor recurrence rates for BCS in 2003–2006 by hospital. b Mastectomy: funnel plot with Ipsilateral breast tumor recurrence rates for mastectomy in 2003–2006 by hospital
10%
5%
0% 0
100
200
300
400
500
600
700
600
700
Number of patients with BCS as final surgery per hospital 95% Cl
b
BCS
20%
IBTR rate
15%
10%
5%
0% 0
100
200
400
300
500
Number of patients with mastectomy as final surgery per hospital 95% Cl
Mastectomy
TABLE 3 Multivariate analyses of hospital characteristics associated with ipsilateral breast tumor recurrence in the Netherlands BCS, adjusted model Aa
BCS, adjusted model B
Mastectomy, adjusted model A
Mastectomy, adjusted model B
HR
HR
95 % CI
HR
HR
95 % CI
95 % CI
95 % CI
b
Hospital volume 2003–2006 Quintile 1
1 (ref)
Quintile 2
1.14
0.83–1.55
1.11
1 (ref) 0.81–1.51
0.79
1 (ref) 0.61–1.04
1 (ref) 0.77
0.59–1.01
Quintile 3
1.32
0.97–1.80
1.32
0.97–1.80
0.87
0.67–1.13
0.84
0.65–1.10
Quintile 4
1.21
0.85–1.74
1.28
0.90–1.84
0.70
0.52–0.94#
0.69
0.51–0.92#
Quintile 5
1.17
0.81–1.69
1.29
0.89–1.87
0.76
0.57–1.02
0.73
0.55–0.98#
Hospital type General
1 (ref)
Teaching
1.03
0.80–1.34
0.98
1 (ref) 0.76–1.28
0.94
1 (ref) 0.77–1.15
1 (ref) 0.94
0.77–1.14
Academic
0.92
0.65–1.31
0.86
0.60–1.23
0.93
0.68–1.27
0.93
0.68–1.28
a
Model A included adjustment for the following variables (as described in Table 1 unless otherwise specified): year of diagnosis, age, T (T2 and T3 combined), N (N2 and N3 combined), grade, histological subtype, multifocality, receptor status, HER2 status and residue; model B included adjustment model A and radiation therapy, chemotherapy and hormonal treatment b
Divided into quintiles based on the total number of specified surgeries per hospital over the period 2003–2006. BCS: 35–167, 168–229, 230–340, 341–425, 426–677; mastectomy: 47–128, 129–209, 210–267, 277–340, 341–543
#
Significant
correction is problematic, especially given the multiple predictors needed. We conclude that this is not a valid indicator to judge the overall quality of breast cancer care of an individual hospital. Another consideration is whether a follow-up of 5 years is long enough to detect all IBTR. Approximately threequarters of all IBTR are diagnosed within 5 years after initial diagnosis.16 Late recurrences (after 5 years) are
more often new, primary tumors, which are associated with better prognosis.17,18 However, the indicator aims to measure the optimal local control of the diagnosed cancer, so true recurrences (mostly occurring within 5 years) are of more interest than second primary breast cancers. The major strength of our study is the availability of a large population based cohort of all breast cancer patients diagnosed in 4 consecutive years in all 92 Dutch hospitals.
Hospital Variation in Breast Tumor Recurrences
The data used in this study was collected by trained registrars and based on a strict coding manual. In case of doubt, registrars were instructed to consult with a registration expert, which minimizes misclassification. In concluding that the IBTR rate is not a suitable outcome indicator for comparing the quality of care between hospitals, the question arises of which indicators do? Process indicators could be of help. A focus on the violation of agreed standards in specific patients may minimize the need for case mix adjustment. These indicators can be quickly available and also pinpoint the target for quick improvements.5,19 The nationwide clinical breast cancer audit in the Netherlands comprises 30 quality indicators, including process indicators (e.g., measuring the proportion of new diagnoses that are discussed in a preoperative multidisciplinary team meeting and the proportion of incomplete first surgeries). The first promising results of this audit were recently presented at the 37th Annual San Antonio Breast Cancer Symposium.20,21 Our findings show that IBTR rates are low in the Netherlands and have substantially improved in recent years. This outcome indicator is of interest on a national level for international comparison but is not suitable for comparing hospital quality of care, because the number of treated patients per hospital is limited and the number of recurrences is too small in individual hospitals. CONFLICT OF INTEREST no conflict of interest.
9.
10.
11.
12. 13.
14.
15.
16.
The authors declare that they have 17.
REFERENCES 1. Birkmeyer JD, Dimick JB, Birkmeyer NJ. Measuring the quality of surgical care: structure, process, or outcomes? J Am Coll Surg. 2004;198(4):626–32. 2. Donabedian A: Evaluating the quality of medical care. Milbank Q. 1966;83(4):691–729. 3. Rubin HR, Pronovost P, Diette GB. From a process of care to a measure: the development and testing of a quality indicator. Int J Qual Health Care. 2001;13(6):489–96. 4. Brien SE, Dixon E, Ghali WA. Measuring and reporting on quality in health care: a framework and road map for improving care. J Surg Oncol. 2009;99(8):462–6. 5. Del Turco MR, Ponti A, Bick U, Biganzoli L, Cserni G, Cutuli B, Decker T, Dietel M, Gentilini O, Kuehn T, et al. Quality indicators in breast cancer care. Eur J Cancer. 2010;46(13):2344–56. 6. Rutgers EJ, Wittenberg J, Kuijpers AC. Kwaliteit van zorg rond mammacarcinoom in de etalage. In: Utrecht: CBO, Orde van Medisch Specialisten; 2006. 7. Rutgers EJ, Group EC. Quality control in the locoregional treatment of breast cancer. Eur J Cancer. 2001;37(4):447–53. 8. van der Heiden-van der Loo M, Ho VK, Damhuis RA, Siesling S, Menke MB, Peeters PH, Rutgers EJ: Percentage of local recurrence following treatment for breast cancer is not a suitable
18.
19.
20.
21.
performance indicator. Nederlands Tijdschrift Geneeskunde. 2010;154:A1984. Schouten LJ, Hoppener P, van den Brandt PA, Knottnerus JA, Jager JJ. Completeness of cancer registration in Limburg, The Netherlands. Int J Epidemiol. 1993;22(3):369–76. Sobin LH, Wittekind C. TNM Classification of malignant tumours. International Union against cancer (UICC), vol. 6. New York: Wiley-Liss; 2002. Moossdorff M, van Roozendaal LM, Strobbe LJ, Aebi S, Cameron DA, Dixon JM, Giuliano AE, Haffty BG, Hickey BE, Hudis CA, et al. Maastricht delphi consensus on event definitions for classification of recurrence in breast cancer research. J Natl Cancer Inst. 2014;106(12):dju288. Spiegelhalter DJ. Funnel plots for comparing institutional performance. Stat Med. 2005;24(8):1185–202. Clarke M, Collins R, Darby S, Davies C, Elphinstone P, Evans E, Godwin J, Gray R, Hicks C, James S, et al. Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet. 2005;366(9503):2087–106. Early Breast Cancer Trialists’ Collaborative G, Darby S, McGale P, Correa C, Taylor C, Arriagada R, Clarke M, Cutter D, Davies C, Ewertz M, et al. Effect of radiotherapy after breast-conserving surgery on 10-year recurrence and 15-year breast cancer death: meta-analysis of individual patient data for 10,801 women in 17 randomised trials. Lancet. 2011;378(9804):1707–16. Kolfschoten NE, Marang van de Mheen PJ, Gooiker GA, Eddes EH, Kievit J, Tollenaar RA, Wouters MW, Dutch Surgical Colorectal Audit Group. Variation in case-mix between hospitals treating colorectal cancer patients in the Netherlands. Eur J Surg Oncol. 2011;37(11):956–63. Park CC, Mitsumori M, Nixon A, Recht A, Connolly J, Gelman R, Silver B, Hetelekidis S, Abner A, Harris JR, et al. Outcome at 8 years after breast-conserving surgery and radiation therapy for invasive breast cancer: influence of margin status and systemic therapy on local recurrence. J Clin Oncol. 2000;18(8):1668–75. Moran MS, Haffty BG. Local-regional breast cancer recurrence: prognostic groups based on patterns of failure. Breast J. 2002;8(2):81–7. Panet-Raymond V, Truong PT, McDonald RE, Alexander C, Ross L, Ryhorchuk A, Watson PH. True recurrence versus new primary: an analysis of ipsilateral breast tumor recurrences after breast-conserving therapy. Int J Radiat Oncol Biol Phys. 2011;81(2):409–17. Lilford R, Mohammed MA, Spiegelhalter D, Thomson R. Use and misuse of process and outcome data in managing performance of acute medical care: avoiding institutional stigma. Lancet. 2004;363(9415):1147–54. Tjan-Heijnen VC, Van Bommel ACM, van der Heiden-van der Loo M, Westenend PJ, De Vries B, Smorenburg CH, Jager A, Lobbes MB, Pijnappel RM, Maduro JH, et al. Multidisciplinary breast cancer care registry and quality control system in the Netherlands: The NABON breast cancer audit. In 37th Annual San Antonio Breast Cancer Symposium. San Antonio;2014. Van Bommel ACM, Vrancken Peeters MJ, van der Heiden-van der Loo M, Van Dalen T, Rutgers E, Wouters MW, Lobbes MB, Pijnappel RM, Mureau MA, Westenend PJ, et al. The NABON breast cancer audit; quality improvement in three years’ time. In 27th Annual San Antonio Breast Cancer Symposium. San Antonio;2014.
