OUP user menu

Timeliness of cancer care from diagnosis to treatment: a comparison between patients with breast, colon, rectal or lung cancer

Xue Li, Andrew Scarfe, Karen King, David Fenton, Charles Butts, Marcy Winget
DOI: http://dx.doi.org/10.1093/intqhc/mzt003 197-204 First published online: 24 January 2013

Abstract

Objective The purpose of this study was to assess the value in measuring specific time intervals across cancer sites to identify potentially important variation in the timeliness of cancer care that may inform needed changes and/or improvements in coordination of care.

Design Retrospective population-level study. Demographic and treatment information were obtained from the Alberta Cancer Registry. Date of oncologist-consult was obtained from cancer medical records.

Setting Alberta, Canada.

Participants All patients diagnosed in 2005 with breast, colon, rectal or lung cancer who were residents of Alberta, Canada.

Main Outcome Measures (i) Number of days from diagnosis to first treatment by treatment modality and cancer site, (ii) number of days from surgery to post-surgery consultation and subsequent treatment and (iii) relationship between clinical and demographic factors and the cancer-specific provincial median time for outcome measures (i) and (ii).

Results Time from diagnosis to surgery, if first treatment, was ∼4 months for lung cancer compared with 1–2 months for breast and colorectal cancers. Factors associated with this time interval for breast and colorectal cancers was stage at diagnosis but was region of residence for lung cancer.

Conclusions Important variation within and across cancer sites identified in the care intervals evaluated in this study provides relevant information to inform local areas for improvement. Comparisons of these intervals across healthcare systems may also provide insights into strengths of different models for coordinating care.

  • health services research
  • lung neoplasm
  • colorectal neoplasm
  • breast neoplasm
  • quality of healthcare
  • quality indicators

Introduction

Timely access is a key dimension to evaluate and monitor quality of care as a priority for improvement in the healthcare system [13]. Various approaches have been taken by countries to tackle issues related to timely access to cancer care [4, 5]. ‘Wait time guarantees’ [5] became a national approach and remain a priority in the Canadian healthcare system [6]. In spite of the efforts, standard evidence-based measures (and corresponding targets) do not exist that can be used to evaluate the quality of care across the cancer care continuum with respect to the timing of healthcare service delivery [79]. One solution that has been proposed is to conduct large population-based studies, to apply the same measures consistently and to look for variation across groups and/or time [10].

Very few studies, however, can be found that have applied measures to assess timely care consistently over time at a population level, and those that have, typically have done so under research conditions that enable collection of data that are not routinely available. Some studies [1113] conducted in Canada, for instance, collected detailed time points such as disease detection, referral to specialist, first specialist visit, receipt of pathology report and treatment decision to evaluate wait times for cancer care. Routine collection of such time points and monitoring of corresponding time intervals between care steps is costly and cannot be performed without large patient chart reviews, involvement of a variety of care providers and/or a robust electronic medical record that links events across providers. Even if the latter is available, it is difficult to ensure consistency in definition and accuracy of collection of a time point such as ‘decision to treat’ across care providers.

A quality healthcare system, however, should measure and evaluate how long it takes patients to travel between key points of care as a fundamental aspect of providing patient-centered care and to ensure healthcare that is actually ‘caring’; the information is also relevant for identifying system (in)efficiencies that may impact both quality of care and costs to the healthcare system, either positively or negatively. Large or unexpected variation across cancer types for a given care interval should give providers and/or administrators cause to investigate the reason(s) to inform whether interventions are needed.

Key points of cancer care include dates of diagnosis, first treatment, surgery and adjuvant treatment. The following two time intervals, (i) time from diagnosis to first treatment and (ii) time from surgery to first post-surgical treatment, are, therefore, operationally meaningful measures of quality that are potentially modifiable and relevant as high-level measures for evaluating the healthcare system. Furthermore, the points of care are defined similarly and are routinely collected by many cancer registries around the world. Comparisons of timeliness of care using these intervals could potentially be made nationally and internationally as have been done for incidence, prevalence and survival.

Breast, colorectal and lung cancers are three of the four most common and prevalent cancers [1416]. The different maturity of screening programs, levels of public and physician awareness and likely varying degree of coordination of care across and within healthcare providers may translate into important differences in the timeliness of care received by these different patient populations [17]. The purpose of this study, therefore, was to assess the value in measuring the above-mentioned time intervals by comparing across tumor sites to identify potentially important variation.

Methods

Study population

All residents of Alberta, Canada, who were diagnosed in 2005 with invasive breast (ICD-O [18] codes: c50.0–c50.6, c50.8 and c50.9), colon (ICD-O codes: c18.0, c18.2–c18.9), rectal (ICD-O codes: c19.9 and c20.9) or lung (ICD-O codes: c34.0–c34.3, c34.8 and c34.9) cancers and treated in the province were identified from the Alberta Cancer Registry. In order to ensure clinical relevance of patients within each tumor site, only adenocarcinoma histologies were included for colorectal cancers and only non-small cell histologies were included for lung cancers. Additionally, patients were excluded if they were: (i) treated outside of the province; (ii) diagnosed with another cancer within the preceding or subsequent 6 months of their diagnosis (may interfere with expected treatment); (iii) died within 1 week of their diagnosis (no opportunity to receive expected treatment) or (iv) did not have a histology that could be staged.

Data source for clinical and demographic information

Demographic and treatment information including patient age and region of residence at diagnosis, date and stage of diagnosis, initial treatment modalities and the start dates of each were obtained from the Alberta Cancer Registry. Cancer stage was determined using the American Joint Committee on cancer staging rules [19]. Initial treatment includes all treatment modalities given at the time of diagnosis to the primary tumor site; it does not include treatment given for disease progression or treatment given to metastatic sites. Date of consultation after surgery was abstracted from the cancer electronic medical records.

Data analysis

The following time intervals were the primary outcome measures of the study: (i) time from diagnosis to first treatment and (ii) time from surgery to post-surgical treatment. Time from surgery to consultation with an oncologist was a secondary interval of interest. Since the impact of the timing of hormone therapy on patient outcomes is not well established, the intervals examined only included surgery, chemotherapy or radiation therapy as treatments relevant for the time intervals.

According to treatment guidelines that existed in 2005 [2023], patients with stage I–IV breast, stage III–IV colon, stage II–IV rectal and stage IB–IV lung cancers should or may be treated with post-surgical chemotherapy and/or radiation therapy. Only patients with these stages of disease at diagnosis were, therefore, included in the analyses evaluating post-surgical care.

Estimates for the 50th and 90th percentiles were calculated for each time interval by cancer site. The inverse Kaplan–Meier [24] graphs were generated for the time interval ‘time from diagnosis to first treatment’ overall and by first treatment modality for each cancer site. The log-rank [25] test was used to determine statistical differences in curves across tumor sites and disease stages.

In order to assess potential sources of variation within a cancer site and determine if associations varied across cancer sites, age, stage and region of residence at diagnosis were examined. Specifically, the time intervals ‘time from diagnosis to first treatment’ and ‘time from surgery to first post-surgery treatment’ were dichotomized based on the cancer-specific provincial median time. The proportion of patients for each variable of interest that received treatment within the cancer-specific provincial median time was calculated. The χ2 or Fisher's exact tests, as appropriate, were used to assess statistical significance. Age was categorized using arbitrary cutoff points of 50, 60, 70 and 80 years at diagnosis. For regional comparisons, patient residence at diagnosis was categorized into five geographical areas: North, Edmonton and area, Central, Calgary and area and South. These align with specific healthcare ‘zones’ in the province. The Northern, Central and Southern regions comprise rural and/or remote regions that have small cancer facilities that provide consultations and/or chemotherapy close to home. Edmonton and Calgary are large urban cities that have the only two tertiary cancer centers in the province and provide oncology consultations, chemotherapy and radiation therapy.

Multiple logistic regression was also conducted to calculate adjusted odds ratios (ORs) to assess the association of age, stage and region of residence by tumor site of receiving treatment within the provincial median time for each cancer site. ORs and 95% confidence intervals (CI) were calculated. All analyses were performed using statistical software SAS 9.1.3 (SAS Institute, Cary, NC, USA) and STATA/SE 10.0 (StataCorp LP, TX, USA).

Results

There were 4636 residents of Alberta who were diagnosed with invasive breast, colon, rectal or non-small cell lung cancer and treated within the province in 2005. The following numbers of patients were excluded for each exclusion criterion: (i) 74 colorectal patients had a histology other than adenocarcinoma; (ii) 142 patients had another cancer within 6 months of diagnosis (may interfere with expected treatment and timing of it); (iii) 74 patients died within a week of diagnosis (no opportunity to receive expected treatment) and (iv) 181 patients did not have a histology that could be staged. A total of 4165 patients were included in the study: 1673 with breast cancer, 807 with colon cancer, 463 with rectal cancer and 1222 with non-small cell lung cancer.

Table 1 presents the distribution of first treatment modalities received by cancer site and summarizes the time from diagnosis to first treatment by cancer site and first treatment modality. Only 64% of patients with lung cancer received first treatment compared with 94% or more of patients with breast, colon or rectal cancer. The first treatment was surgery for 90% of patients with breast or colon cancer, 67% of patients with rectal cancer and only 18% of patients with lung cancer. Radiation was the most common treatment modality first received by patients with lung cancer. For those whose first treatment was surgery, 90% of breast, colon, rectal and lung cancer received it within ∼2, 1, 2 and 4 months of their diagnosis, respectively.

View this table:
Table 1

Time from diagnosis to first treatment by treatment modality and cancer site

Cancer siteDiagnosis (n)1st treatmentn (%)aNumber of days
Median90%
Breast1673Overall1614 (96)2963
 Surgery1508 (90)2862
 Other106 (6)3278
  Chemotherapy99 (6)3178
  Radiation therapy7 (<1)3493
Colon807Overall763 (94)439
 Surgery727 (90)335
 Other36 (4)5790
  Chemotherapy36 (4)5790
Rectal463Overall444 (96)3075
 Surgery310 (67)2168
 Other134 (29)4677
  Chemotherapyb94 (20)5079
  Radiation therapy40 (9)4276
Lung1222Overall785 (64)41115
 Surgery219 (18)49129
 Other566 (46)39107
  Chemotherapyb144 (11)52111
  Radiation therapy422 (35)35104
  • aDenominator for percent is based on all diagnosed patients of the site. bCombined chemo-radiation therapy n = 79 for rectal cancer and n = 18 for lung cancer.

Figure 1 displays the cumulative time from diagnosis to first treatment by treatment type and cancer site. The time from diagnosis to first treatment overall and when the first treatment was surgery or chemotherapy varies significantly by cancer sites (P < 0.0001 for all, Fig 1a–c). There is not a significant difference in the timing of treatment between rectal and lung cancer patients if the first treatment was radiation therapy or combined chemo-radiation therapy (P = 0.90, Fig. 1d). Patients with lung cancer experienced the longest elapsed time from diagnosis to first treatment (Fig. 1a), while colon cancer patients had the shortest elapsed time; 43% of colon cancer patients were diagnosed at the time of their surgery (Fig. 1b). Patients with breast cancer received chemotherapy sooner after diagnosis than those with other tumor types (P < 0.0001, Fig. 1c).

Figure 1

Cumulative time from diagnosis to first treatment by treatment modality and cancer site.

Table 2 presents the number of patients who had surgery followed by a consultation with an oncologist and treatment with chemotherapy or radiation therapy; time to each event from surgery by cancer site is also presented. The median and 90th percentile time from surgery to consultation each vary by ∼1 week: the median ranges from 39 to 47 days and the 90th percentile ranges from 62 to 69 days (Table 2). Similarly, in Table 2, the variation in the median and 90th percentile for the time from surgery to subsequent treatment by tumor site is 6 and 13 days, respectively: the median ranges from 57 to 63 days and the 90th percentile ranges from 89 to 104 days.

View this table:
Table 2

Time from surgery to post-surgery consultation and subsequent treatment

Cancer siteSurgery (n)Post-surgical eventn (%)aNumber of days
Median90%
Breast
 Stage I–IV1579Consultation with an oncologist1455 (92)4062
Chemotherapy or radiation therapy1025 (65)5790
Colon
 Stage III–IV329Consultation with an oncologist286 (87)3969
Chemotherapy or radiation therapy192 (58)63104
Rectal
 Stage II–IV301Consultation with an oncologist250 (83)4066
Chemotherapy or radiation therapy180 (60)63103
Lung
 Stage IB–IV151Consultation with an oncologist123 (81)4769
Chemotherapy or radiation therapy84 (56)5889
  • aDenominator for percent is based on all patients who had surgery for each tumor site.

Tables 3 and 4 show the results of the unadjusted and adjusted analyses of receiving treatment within the provincial site-specific median time from diagnosis to first treatment by age, stage and region of residence at diagnosis. Results are presented separately by whether the first treatment was surgery (Table 3) or chemotherapy/radiation therapy (Table 4). Only results for rectal and lung cancers are shown in Table 4 due to the small number of patients with breast or colon cancer who received chemotherapy or radiation therapy as a first treatment.

View this table:
Table 3

Unadjusted and adjusted estimates of associations with time from diagnosis to first treatment (surgery) within the cancer-specific provincial median time

CharacteristicsBreastColonRectalLung
n (%)aORb (95% CI)n (%)aORb (95% CI)n (%)aORb (95% CI)n (%)aORb (95% CI)
Total755 (50)375 (52)156 (50)110 (50)
AgeP* = 0.39P* = 0.36P* = 0.18P* = 0.96
 <50205 (52)Reference
 50–59191 (51)1.0 (0.8–1.4)90 (53)cReferencec48 (56)cReferencec27 (51)cReferencec
 60–69164 (51)1.0 (0.8–1.4)84 (50)0.9 (0.6–1.3)44 (48)0.6 (0.3–1.2)41 (51)1.1 (0.5–2.2)
 70–79124 (48)0.9 (0.7–1.4)107 (48)0.9 (0.6–1.3)35 (43)0.6 (0.3–1.0)42 (49)d1.1 (0.5–2.3)d
 ≥8071 (45)0.7 (0.5–1.1)94 (56)1.2 (0.8–1.9)29 (57)1.0 (0.5–2.1)
StageP* < 0.0001P* = 0.002P = 0.002P* = 0.58
 Stage I341 (45)Reference58 (43)Reference38 (36)Reference81 (54)Reference
 Stage II299 (55)1.6 (1.3–2.0)136 (51)1.4 (0.9–2.1)46 (58)2.5 (1.4–4.6)14 (41)0.7 (0.3–1.5)
 Stage III103 (59)1.9 (1.3–2.6)90 (48)1.2 (0.8–1.9)50 (54)2.1 (1.2–3.8)12 (41)0.7 (0.3–1.6)
 Stage IV12 (50)1.3 (0.6–2.9)91 (65)2.5 (1.5–4.1)22 (67)4.0 (1.7–9.3)ee
RegionP* = 0.35P* = 0.16P* = 0.60P* = 0.03
 Edmonton251 (51)Reference120 (47)Reference47 (51)Reference35 (42)Reference
 Northern75 (54)1.1 (0.7–1.6)39 (47)1.0 (0.6–1.6)16 (50)0.9 (0.4–2.0)13 (39)1.0 (0.4–2.2)
 Central101 (52)1.0 (0.7–1.4)57 (59)1.5 (1.0–2.5)30 (61)1.4 (0.7–2.9)10 (45)1.2 (0.5–3.1)
 Calgary264 (47)0.8 (0.6–1.1)132 (56)1.4 (1.0–2.0)52 (47)0.8 (0.5–1.4)44 (68)2.9 (1.5–5.8)
 Southern64 (55)1.1 (0.8–1.7)27 (47)1.0 (0.5–1.7)11 (42)0.7 (0.3–1.8)8 (53)1.6 (0.5–4.9)
  • aPercent was based on the total number of patients who received first treatment if it is surgery in each category.bOdds ratio: relative to the reference category, the table reports the odds that patients in the category received surgery within the median provincial time from diagnosis, adjusted for all the factors listed in the table for each outcome variable. A value of >1.0 indicates higher odds of receiving surgery within the provincial median time than the reference group.cAge group is <60.dAge group is ≥70.eResults for sample size ≤5 in the cell were not shown.

  • *P-value: the probability of observing what we observed (or more extreme) if the characteristic is not associated with time to first treatment.

View this table:
Table 4

Unadjusted and adjusted estimates of associations with time from diagnosis to first treatment (non-surgical) within the cancer-specific provincial median time

CharacteristicsRectalLung
n (%)aORb (95% CI)n (%)aORb (95% CI)
Total72 (54)284 (50)
AgeP* = 0.19P* = 0.12
 <6031 (62)Reference79 (58)Reference
 60–6919 (56)1.1 (0.4–3.0)82 (51)0.7 (0.4–1.1)
 ≥7022 (44)0.5 (0.2–1.3)123 (46)0.6 (0.4–1.0)
StageP* = 0.09P* < 0.0001
 Stage I0 (0)8 (17)Reference
 Stage II20 (43)Referencecc
 Stage III31 (61)2.7 (0.9–7.6)79 (42)3.5 (1.5–8.1)
 Stage IV21 (58)0.4 (1.0–6.6)193 (60)8.2 (3.6–18.7)
RegionP* = 0.01P* = 0.02
 Edmonton29 (52)Reference111 (58)Reference
 Northern12 (92)29.0 (3.0–280.3)30 (64)1.5 (0.7–3.1)
 Central8 (36)0.7 (0.2–2.4)34 (40)0.5 (0.3–0.9)
 Calgary23 (64)2.8 (1.0–7.7)90 (44)0.6 (0.4–1.0)
 Southern0 (0)19 (53)0.9 (0.4–2.0)
Treatment typeP* = 0.05P* < 0.0001
 Chemotherapy49 (52)Reference47 (33)Reference
 Radiation therapy23 (58)2.7 (1.0–7.1)237 (56)4.0 (2.6–6.2)
  • aPercent was based on the total number of patients who received first treatment if it is chemotherapy or radiation therapy in each category.bOdds ratio: relative to the reference category, the table reports the odds that patient in the category received non-surgical treatment within the median provincial time from diagnosis, adjusted for all the factors listed in the table for each outcome variable. A value of >1.0 indicates higher odds of receiving non-surgical treatment within the provincial median time than the reference group. The results were not abstracted for breast and colon cancer due to small sample size.cResults for sample size ≤5 in the cell were not shown.

  • *P-value: the probability of observing what we observed (or more extreme) if the characteristic is not associated with time to first treatment.

The main results of interest shown in these tables are: (i) time to first treatment does not vary by age for any of the cancer sites regardless of first treatment modality; (ii) stage is generally associated with time from diagnosis to first treatment but there are differences in the associations across cancer sites and by first treatment and (iii) region of patient residence is an important factor related to time to first treatment for lung cancer patients but does not appear to be for the other cancers examined.

As mentioned, there is a strong association between stage and time from diagnosis to surgery for breast, colon and rectal cancer patients (P < 0.0001, 0.002 and 0.002, respectively, Table 3). Based on the 95% CI, the relationship across stage varied by each of these tumor sites and was generally not an ordinal relationship. Specifically, patients with stage II and III breast cancer, stage IV colon cancer or stage II–IV rectal cancer were more likely to receive surgery within the respective provincial site-specific median times from diagnosis than their counterparts. In contrast, although time from diagnosis to surgery did not vary by stage for lung cancer patients, it did if the first treatment was chemotherapy or radiation therapy (Table 4). Lung cancer patients with stage III or IV disease were more likely to receive chemotherapy or radiation therapy within the provincial median time [OR = 3.5 (95% CI: 1.5–8.1); OR = 8.2 (95% CI: 3.6–18.7), respectively] than stage I lung cancer patients.

Region of patient residence was associated with the median time from diagnosis to first treatment for lung cancer patients regardless of what their first treatment was (adjusted P = 0.03 Table 3 and adjusted P = 0.02 Table 4) but for rectal cancer patients, was only associated with the median time from diagnosis to chemotherapy or radiation therapy (adjusted P = 0.01 Table 4). The regional differences for lung cancer patients were largely due to differences observed between the regions with the largest cities. Specifically, relative to lung cancer patients living in Edmonton, those living in Calgary were more likely to receive surgery within the provincial median time from diagnosis [OR = 2.9 (95% CI: 1.5–5.8)], whereas they were less likely to receive chemotherapy or radiation therapy within the provincial median time from diagnosis [OR = 0.6 (95% CI: 0.4–1.0)].

Also shown in Table 4, both rectal and lung cancer patients who received radiation therapy as their first treatment were more likely to receive it within the cancer-specific provincial median time than those who received chemotherapy as their first treatment [OR = 2.7 (95% CI: 1.0–7.1) and OR = 4.0 (95% CI: 2.6–6.2), respectively].

Table 5 shows the results of the unadjusted and adjusted analyses of receiving post-surgical treatment within the cancer-specific provincial median time from surgery by age, stage and region of residence at diagnosis. Unlike the results of the analyses of time from diagnosis to first treatment, patient age was associated with time from surgery to post-surgical treatment for breast and lung cancers, region of residence was associated with breast, rectal and lung cancer patients and treatment type was associated with breast cancer but not rectal or lung cancer. Similar to the results of the analyses of time from diagnosis to first treatment, stage at diagnosis was associated with time from surgery to post-surgical treatment for breast, colon and rectal cancer but not lung cancer.

View this table:
Table 5

Unadjusted and adjusted estimates of associations with time from surgery to post-surgery treatment within the cancer-specific provincial median time

CharacteristicsBreastColonRectalLung
Stage I–IVStage III–IVStage II–IVStage IB–IV
n (%)aORb (95% CI)n (%)aORb (95% CI)n (%)aORb (95% CI)n (%)aORb (95% CI)
Total526 (51)106 (55)94 (52)47 (56)
AgeP* = 0.001P* = 0.80P* = 0.24P* = 0.05
 <50228 (64)Reference
 50–59152 (50)0.6 (0.4–0.9)41 (56)cReferencec38 (54)cReferencec24 (73)cReferencec
 60–6997 (43)0.5 (0.4–0.7)26 (51)0.8 (0.4–1.7)36 (57)1.6 (0.7–3.3)18 (50)0.3 (0.1–1.0)
 70–7936 (36)0.5 (0.3–0.8)39 (57)d1.0 (0.5–2.0)d20 (43)d0.8 (0.3–1.7)dd,ed,e
 ≥8013 (33)0.4 (0.2–0.9)
StageP* = 0.01P* = 0.0003P* = 0.001P* = 0.58
 Stage I185 (42)Reference28 (62)Reference
 Stage II212 (55)0.9 (0.7–1.3)30 (50)Reference11 (61)0.8 (0.2–2.9)
 Stage III124 (65)1.8 (1.2–2.7)75 (66)Reference60 (62)1.7 (0.8–3.3)7 (39)0.4 (0.1–1.5)
 Stage IVee31 (39)0.3 (0.2–0.6)eeee
RegionP* = 0.03P* = 0.55P* = 0.03P* = 0.07
 Edmonton183 (54)Reference40 (56)Reference38 (64)Reference11 (39)Reference
 Northern55 (56)1.0 (0.6–1.6)9 (45)0.6 (0.2–1.7)14 (61)0.9 (0.3–2.6)7 (54)2.9 (0.7–12.8)
 Central70 (58)1.0 (0.6–1.5)10 (45)0.7 (0.2–1.8)14 (50)0.5 (0.2–1.3)e1.0 (0.2–5.1)
 Calgary185 (46)0.6 (0.5–0.9)36 (57)1.0 (0.5–2.0)23 (38)0.3 (0.1–0.6)21 (78)5.2 (1.5–18.3)
 Southern33 (49)0.8 (0.4–1.3)11 (69)1.8 (0.6–6.1)eeee
Treatment TypeP* < 0.0001
 Chemotherapy370 (63)Reference106 (55)93 (53)47 (57)
 Radiation therapy156 (36)0.4 (0.3–0.5)ee
  • aPercent was based on the total number of patients who received post-surgery treatment (chemotherapy or radiation therapy) in each category.bOdds ratio: relative to the reference category, the table reports the odds that patient in the category received post-surgical treatment within the median provincial time from surgery, adjusted for all the factors listed in the table for each outcome variable. A value of >1.0 indicates higher odds of receiving post-surgical treatment within the provincial median time than the reference group.cAge group is <60.dAge group is ≥70.eResults for sample sizes ≤5 in the cell were not shown.

  • *P-value: the probability of observing what we observed (or more extreme) if the characteristic is not associated with time to first treatment.

The relationship between age and time from surgery to post-surgical treatment was similar for both breast and lung cancer patients; as patient age increased, there was a decreased likelihood of being treated within the cancer-specific provincial median time (adjusted P = 0.001 and 0.05, respectively). Similar to the relationship between stage at diagnosis and time from diagnosis to first treatment, the relationship between stage and time from surgery to post-surgical treatment varied by site and was not generally ordinal.

A common factor of the variation by region for time from surgery to post-surgical treatment was that the primary difference was due to differences between the two urban centers, as found with the time interval diagnosis to first treatment. Breast and rectal cancer patients living in Calgary were less likely to receive post-surgical treatment within the cancer-specific provincial median time [OR = 0.6 (95% CI: 0.5–0.9) and OR = 0.3 (95% CI: 0.1–0.6), respectively] relative to patients living in Edmonton; however, the opposite was true for lung cancer patients [OR = 5.2 (95% CI: 1.5–18.3)].

Discussion

The purpose of the study was to assess the value of using the high-level measures of ‘time from diagnosis to first treatment’ and ‘time from surgery to post-surgical treatment’ to identify potentially important variation in a healthcare system that can and should be addressed. Based on the results of this study, we conclude that the proposed measures are practical and relevant for identifying potentially important variation within and across cancer sites. Evaluation of these intervals by first treatment modality, age, stage and region of residence identified potential issues related to coordination of healthcare, referral and treatment patterns, and patient populations ‘at risk’ for suboptimal care. Although the results presented are specific to Alberta, Canada, the methods are transferable and relevant to any healthcare system. Furthermore, the data needed for analyses as presented herein are relatively available and of high quality in many if not most healthcare systems making the methodology feasible.

A finding that might be generalizable to many healthcare systems was that although there were many differences across the cancer sites with respect to the timing of events and variables associated with the time, lung cancer tended to be the most unique. For instance, time from diagnosis to surgery for lung cancer patients was ∼4 months rather than 1–2 months for the other cancers. Unlike the other cancer sites, large regional variation in the timing of treatment, relative to the median time, existed for lung cancer patients in all time intervals evaluated in both adjusted and unadjusted analyses. Regional variation was seen in other cancer sites for certain intervals but was not consistently present as it was the lung cancer.

This suggests that there may be large systemic issues related to coordination of care for lung cancer patients throughout the patient care trajectory, whereas there may be smaller specific issues related to care coordination for the other cancer sites. Other studies have also found that lung cancer patients tend to have many unmet needs at least some of which are due to poor coordination of care [26, 27]. Further examination is needed to identify where specific targeted intervention/effort is needed.

In conclusion, we have found that important variation in cancer care can be identified by comparing time from diagnosis to first treatment and time from surgery to first post-surgical treatment between and across cancer sites. This information is valuable locally in identifying areas to target for improvement and also to assess the impact of interventions implemented. International or cross-jurisdictional comparisons across healthcare systems with different delivery models of care may provide insights into strengths of different models for coordinating care, thus enhancing current and future targeted efforts to improve the quality of cancer care.

Funding

This work was supported by the Alberta Cancer Foundation (grant number 23932).

Acknowledgements

The authors thank Angela Bella for assistance in the formatting of tables, figures and references of this manuscript.

References

View Abstract