The 30-day hospital readmission and mortality after surgery in colorectal cancer patients

Colorectal cancer (CRC) is the most diagnosed cancer in Saudi males and the third most common cancer in females, with a 5-year observed survival rate of 52% [1,2]. Given the improved outcomes associated with the surgical resection of CRC, surgery remains the principal treatment modality for patients diagnosed with CRC [3]. Nonetheless, surgical outcomes including hospital readmission during the first weeks after surgery are prevalent, costly, and implicated with increased mortality [4‐6]. Patients with cancer who have been discharged from hospitals have a readmission rate of up to 27% [7] and constitute a healthcare burden associated with morbidity and mortality [4]. No prior studies assessed the CRC rate of hospital readmissions and the risk factors associated with readmission among the Saudi population.

International studies, however, have evaluated CRC surgical outcomes including the 30-day hospital readmission and its predictors. Several reasons were associated with readmission including age, gender, race/ethnicity, type of health insurance, socioeconomic status (SES), tumor stage, comorbidities, procedure type, surgical approach (open or laparoscopic), length of hospital stay (LOS), surgical complications, non-home discharge, blood transfusion, and stoma creation [8]. Whether the rate and predictors of CRC readmissions are the same in the Saudi population remains unknown. This is vital given the poor characteristics of CRC in the Saudi population such as late-stage at diagnosis and low survival [9].

In Saudi Arabia, the total number of surgeries during the year 2020 was 398,188 and 46,831 of them were implemented at the Ministry of National Guard- Health Affairs (MNG-HA) [10]. More than 80% of the CRC patients at the MNG-HA are treated surgically [11], and hence elucidating the surgical outcomes of these patients is crucial. Similar to the US Veteran Affairs (VA) health system, the MNG-HA is a health system that covers eligible members in all regions of Saudi Arabia. Moreover, at the national level, Saudi Arabia has developed 12 strategic programs including a health sector transformation program that is mainly focusing on improving the population's health through efficient use of resources.

The present study aims to assess the 30-day hospital readmission after index colorectal cancer surgery and evaluates the risk factors that impact hospital readmission. We principally focused on the impact of demographic variables, tumor characteristics including the stage at diagnosis and LOS. We also investigated the reasons for hospital readmission. Identifying the rate of readmission and its predictors will contribute to the development of interventions that should reduce readmission, hospitalization costs and improve cancer patient outcomes.

Figure 1 shows the eligibility criteria for the study population. In the present study, 356 patients were included, and 49 (13.76%) of them were readmitted, while the remaining were non-readmitted patients (Table 1). The average age was 60.79 years with almost 60% of the sample being males and more than 78% were married. The average BMI was 27.28 kg/m² and nearly 70% of the patients did not have any comorbid conditions. Compared to the readmitted patients, non-readmitted individuals were more likely to be male and presented with regional stage at diagnosis. There was a significant association between stage at diagnosis and hospital readmission (P value = 0.03).

Table 2 displays the characteristics of the index surgery stratified by readmission status. The majority of the patients have a tumor located in the colon (83.71%) and the rest were in the rectum. Whereas 69.70% of the patients underwent open surgery, only 30.40% underwent laparoscopic surgery. Most patients had elective surgery (84.55%) and were discharged to the post-anesthesia care unit (82.87%). Furthermore, reasons for hospitalization after index discharge are reported in Table 3 and Additional file 1: Appendix 1. The recorded reasons were gastrointestinal (18.42% colon vs. 36.36% rectum), urinary tract infection (21.05% colon vs. 0% rectum), surgical site infection (13.16% colon vs. 9.09% rectum) and stoma related (2.63% colon vs. 27.27% rectum). Almost a quarter of readmission were due to gastrointestinal causes such as anastomotic leak, bowel obstruction, perforated tumor and rectal bleeding/discharge. Nevertheless, none of the differences were significant.

In the univariate analysis, there was a positive association between age and readmission and females had higher odds of readmission than males, but both did not reach a significance level (Table 4). On the contrary, compared to patients diagnosed with a localized tumor, those diagnosed with distant metastatic disease have 4.09 higher odds of readmission (odds ratio 4.09, 95% confidence intervals 1.29–12.98). Likewise, patients discharged to locations other than post-anesthesia care or ICU have 3.75 higher odds of readmission (odds ratio 3.75, 95% confidence intervals 1.19–11.80).

In the multivariable analysis, both patients diagnosed with distant metastatic tumors and those discharged to locations other than post-anesthesia care or ICU have higher odds of readmission (Table 4). Finally, for the mortality outcome, there is no statistically significant difference in the survivability amongst those readmitted and those not readmitted (P = 0.0581) (Fig. 2).

In this retrospective population-based cohort study, we found that 13.76% of CRC patients were readmitted within 30-day of index hospitalization, a finding that is slightly higher than previous studies (9–11%) [4,12‐17]. The majority of recorded reasons for hospital readmissions were gastrointestinal (22.45%), urinary tract infection (16.33%), and surgical site infection (12.24%). Almost one-quarter of metastatic patients were readmitted within 30 day of index hospitalization.

While previous studies found that 22% to 27% of metastatic CRC patients were readmitted [18‐21,24], we observed that 24.30% (17/70) of our metastatic patients were readmitted. In the multivariable analysis, metastatic patients had 3.86 higher odds of 30-day hospital readmission compared to those with localized disease, similar to previous findings [22,23]. Several studies from Saudi Arabia have shown that a sizable percentage of Saudi CRC patients were diagnosed at an advanced stage [2,9], a risk factor that has been shown to increase readmission in our population. Moreover, we found that patients discharged to locations other than post anesthesia care unit were more likely to be readmitted.

Taken together, hospitalized patients with metastatic disease should be counseled before discharge, for example through outpatient transition [24], to reduce hospital readmission. Alternatively, down-staging efforts through an early stage at diagnosis, as secondary public health prevention, will indirectly reduce the rate of hospital readmission [2,9].

Similar to some but not all prior research, we found no association between readmission and comorbidities, LOS, or surgical approach. In a population-based study that assessed readmission rates in the VA population, the authors found no significant association between the aforementioned factors on the rate of readmission [25]. On the contrary, some other studies found a positive association with comorbidities, LOS, and open surgery. Notably, our population has longer index hospitalization compared to other populations [4,5,13,26]. It is possible that lack of association is due to the small observed number in the readmitted patients particularly those with > one comorbidity score.

The results of the present study should be interpreted within the scope of the following limitations. First, the reported results should be generalized to the MNG-HA population or a similar population. Second, SES is a factor that was affecting readmission in some previous studies which were not accounted for in the current study. Nonetheless, given that the MNG-HA population has equitable access to care and that all members are employed by the system, the SES effect is modest. Third, among patients with metastatic disease, we were not able to distinguish between primary tumor resection from metastatic resection. Fourth, given the positive volume-outcome relationship in CRC patients, adjusting for such factors could have improved the finding of the present study. Lastly, some of our admitted patients (n = 128) were missing discharge date and were excluded from analysis. The characteristics of these patients, nonetheless, were similar to our study population (Additional file 1: Appendix 2).

The knowledge of the rate and factors for hospital readmission in CRC patients has a significant impact on patients and the healthcare system. Given the identified factors, implementing strategies that may reduce readmission rates is needed. For instance, the adoption of minimally invasive surgery (e.g. laparoscopic procedure) could potentially contribute to lower hospitalization after major surgeries such as CRC. Additionally, post-discharge strategies include shorter outpatient follow-up time, nursing or home health care visits, and making a nursing/ educator home phone call. Many other modalities may be studied and implemented.

CRC readmission is common, especially in patients with metastatic disease. Strategies to reduce readmission include planned transition to outpatient care, especially among patients with a high risk of readmission.