Time on wait lists for coronary bypass surgery in British Columbia, Canada, 1991 – 2000

The provincial Cardiac Surgery Registry, a part of BC Cardiac Registries, was created in 1990 to collect data for reporting, planning and research purposes of participating surgeons and hospitals, and the provincial Ministry of Health[11]. The Registry prospectively captures the occurrence and timing of registration, surgery, or removal from the wait lists without surgery, for all patients accepted for cardiac surgery procedures in the four hospitals delivering all adult open-heart surgery services to four million residents of BC.

Between 1991 and 2000, from 15 to 20 cardiac surgeons were performing bypass surgery in BC, with less than 30% turnover. Although cardiac surgeons manage their wait lists independently, they all routinely provide information to the Registry entry modules: surgery registration, operative report, wait-list reconciliation, and discharge summary. When accepting patients on their wait lists, the surgeons document the indication for the procedure as well as the priority for treatment using common criteria (see below).

Once the operation is completed, the operative report containing the procedure and clinical data is entered in the Registry. Patients are removed from the wait lists after undergoing the operation or for other reasons: if they died, declined the operation, accepted surgery from another surgeon, moved away, or switched to medical management. The crude agreement between the Registry and hospital charts for ten demographic and clinical data elements has been estimated at 86%[11].

We deterministically linked the Registry records to administrative databases storing records of all hospital episodes in BC[12]. These records include the dates of admission, procedure and discharge, as well as diagnoses at discharge[13]. These data were used to corroborate the service dates and to identify coexisting medical conditions[14].

If angioplasty is not indicated when the cardiologist evaluates the arterial lesions on the coronary angiogram, then a cardiac surgeon is consulted to assess the patients' suitability for CABG. Patients are transferred to an in-patient ward directly from the catheterization laboratory if expedited assessment is necessary. If deemed suitable, these patients wait for CABG in hospital without registration on a wait list. Alternatively, a consultation with the surgeon can be scheduled at a later date. Surgeons register on their wait lists patients who need CABG and for whom the operation can be safely delayed. As in-patients were not added to wait lists, they were not included in analyses of wait-list times.

There were 9,366 records of registration for isolated CABG added to the Registry between January 1991 and December 2000. We excluded 135 records of patients who were: emergency cases (30), removed on the registration date (101), and had missing operating room reports (4). All remaining 9,231 records had either the surgery date or the date and reason of removal from the list without surgery. We restricted the analyses to the first 52 weeks after registration so that 475 (5%) patients remaining on the lists at 12 months were censored. Of those, 167 eventually underwent surgery; seven died; 78 received medical treatment; 104 declined surgery; 17 were transferred to another surgeon or hospital; and 102 were removed for other reasons.

When assigning priority, all cardiac surgeons in BC apply common guidelines developed in 1990 [see Additional file 1]. Using the location and degree of affected coronary anatomy and symptoms, the guidelines help to: identify patients for whom CABG can increase survival or improve quality of life[15]; classify patients according to urgency of treatment; and assign a maximum recommended waiting time (MRWT). Patients are assigned priority 1 if they require CABG urgently (eg, left main coronary artery stenosis greater than 70%, MRWT three days); priority 2 if there is moderate urgency (eg persistent unstable angina, MRWT six weeks); or priority 3 if there is less urgency (eg intractable chronic angina, MRWT 12 weeks). These guidelines did not undergo any major revisions through the entire period under study.

Using the administrative data, coexisting medical conditions were identified using all primary and secondary discharge diagnoses recorded in all hospital discharge abstracts within one year prior to registration[13]. This time frame was chosen in order to capture the presence of chronic diseases that could have affected the waiting time[14].

For each patient, we identified the presence of major and minor comorbid medical conditions present at registration.

Waiting times were analyzed as prospective observations beginning at the time of registration. Each subject had a wait-list time calculated in calendar weeks from registration to surgery or removal for other reasons. The cumulative probability of undergoing surgery as a function of wait-list time was estimated using the Kaplan-Meier method[16]. Patients removed from the list for reasons other than surgery were treated as censored observations.

Primary comparisons were done across synthetic cohorts of patients defined by two-year periods of registration on the wait lists. Differences in the distributions of wait-list times across cohorts were examined using the log rank-test[17]. The average weekly surgery rate was calculated by dividing the number of operations by the total number of patient-weeks on the list. The effect size for each registration period was estimated by hazard ratios for surgery derived from a Cox proportional hazards model[18]. Hazard ratios (HR) associated with registration periods evaluated the conditional weekly probability of undergoing CABG relative to the 1991–92 period. The priority groups and the presence of comorbidity at registration were included as independent variables in the Cox model to estimate adjusted effects. Age and sex were entered into the regression models as strata variables to avoid the proportionality assumption on these factors while using the proportional hazards model.

The Clinical Research Ethics Board of the University of British Columbia approved the study protocol.

The differences in the proportion of patients undergoing CABG were significant across registration periods (log rank test = 97.3, P < 0.0001), with longer wait-list times for those registered between 1995 and 1998, Figure 2.

Table 3 shows the number of weeks required for a specified proportion of patients to undergo the operation across registration periods. Wait-list times in 1995–96 were such that 10%, 25%, 50%, and 75% patients underwent surgery within 1, 6, 15, and 26 weeks, respectively, whereas half of the 1991–92 cohort underwent surgery within 9 weeks, and 75% did so within 19 weeks. Comparing the 1995–96, 1997–98 and 1999–2000 cohorts we observed a compression in access to surgery, i.e., reduction in the length of wait-list interval required for a specified proportion to undergo the operation. As measured by the difference between 90^th and 50^th percentiles of the wait time distributions, 40% of the 1995–96 cohort underwent surgery within 33 weeks following the median time, while it took 29 weeks for the 1999–2000 cohort.

While the median wait-list time was 11 weeks (the MRWT of priority group 3) in all cohorts combined, 15% of the 1991–92 and 1993–94 cohorts, 22% of the 1995–96 cohort, 19% of the 1997–98 cohort, and 14% of the 1999–2000 cohort experienced an excessive wait, defined as longer than 26 weeks (data not shown). The average weekly number of operations per 100 patients listed varied from 6.5 (6.2–6.9) in the 1991–92 cohort to 5.1 (4.8–5.3) in the 1995–96 cohort to 6.2 (5.9–6.6) in the 1999–2000 cohort, Table 4 (fourth column). Corresponding hazard ratios and 95% confidence intervals (CI) are shown in Table 4, columns 6 and 7. Relative to the 1991–92 cohort, the conditional weekly probabilities of undergoing surgery were 30% lower among 1995–96 patients, HR = 0.70 (0.65–0.76), and 23% lower in 1997–98 patients, HR = 0.77 (0.71–0.83), after adjusting for priority, comorbidity, age and sex. There were no differences between periods 1991–92 and 1999–2000, HR = 0.94 (0.88–1.02).

In each priority group, the proportions of patients undergoing CABG at each week on the wait-list was lower among those registered in 1995–96 compared to 1991–92 as measured by log-rank tests (priority 1: chi-square = 5.6, P = 0.0183, 1 df; priority 2: chi-square = 58.2, P < 0.0001, 1 df; priority 3: chi-square = 20.5, P < 0.0001, 1 df). By 1999–2000, the pattern of change was different between priority groups.

In priority group 1, the average weekly number of operations per 100 patients listed declined from 42.4 (34.6–50.2) in the 1991–92 cohort to 20.3 (17.7–22.9) in the 1995–96 cohort to 12.2 (9.2–15.3) in the 1999–2000 cohort (data not shown). Corresponding HRs and 95% CIs are shown in Table 5, columns 2 and 3. The conditional weekly probabilities of undergoing surgery were 34% lower for the 1995–96 cohort, HR = 0.66 (0.50–0.87), and 53% lower for the 1999–2000 cohort, HR = 0.47 (0.33–0.68), relative to 1991–92. There was a difference in the distribution of wait-list times between 1995–96 and 1999–2000 cohorts (chi-square = 9.9, P = 0.0017, 1 df).

In priority group 2, the average weekly number of operations per 100 patients listed varied from 7.3 (6.8–7.7) in the 1991–92 cohort to 5.3 (5.0–5.6) in the 1995–96 cohort to 7.3 (6.9–7.7) in the 1999–2000 cohort. The adjusted HR in 1999–2000 was 0.99 (0.90–1.08) relative to 1991–92 (Table 5, columns 4 and 5). There was no difference in the distribution of wait-list times between 1991–92 and 1999–2000 cohorts (chi-square = 0.5, P = 0.5, 1 df).

In priority group 3, the average weekly number of operations per 100 patients listed changed from 3.9 (3.4–4.3) in 1991–92 to 2.8 (2.4–3.1) in 1995–96 to 4.0 (3.6–4.5) in 1999–2000. The adjusted HR associated with the 1999–2000 registration period was 1.07 (0.90–1.26) relative to 1991–92 (Table 5, columns 6 and 7). There was no difference between the between 1991–92 and 1999–2000 cohorts (chi-square= 0.6, P = 0.4, 1 df).