Abstract

Objective To evaluate health system use, health outcomes, and avoided costs when patients with chronic gastrointestinal (GI) conditions are managed in the medical home.

Design A retrospective, observational cohort study was conducted through a single-point-of-entry referral system.

Participants Patients with referrals for any of 7 nonurgent indications.

Main outcome measures Patients with referrals for any of 7 nonurgent indications were returned to primary care for management, guided by evidence-based primary care pathways. Patients were linked to administrative databases to extract indications, re-referral rates, endoscopy findings and outcomes, number of emergency department and urgent care visits, and number of hospital admissions. Costs avoided were estimated using Canadian Institute for Health Information data for health care use, consultation, and endoscopy.

Results Between July 1, 2018, and May 31, 2020, a total of 3435 routine referrals were closed for 3274 patients. The most common pathway used was dyspepsia (1154 of 3435, 33.6%). A total of 362 patients (11.1%) had 616 GI-related emergency department or urgent care visits; 52 (1.6%) patients experienced 68 GI-related hospitalizations. A total of 396 patients (12.1%) underwent endoscopy; of the 348 patients with findings available for analysis, 7.8% exhibited a clinically significant finding. The estimated total cost avoided was $1,477,237.

Conclusion The implementation of co-developed primary care pathways safely supports the care of patients with common, nonurgent GI conditions within the medical home. In this population, rates of re-referral and health care use were low, resulting in avoidance of substantial costs and improved overall appropriateness of care.

Gastrointestinal (GI) symptoms due to common, benign conditions are among the most costly to manage.1 Demand for specialty assessment and investigation of patients with GI symptoms is on the rise, and a constrained supply of GI specialists has resulted in limited access to and long wait times for GI specialist care across Canada.2-5

Almost 40% of the general population are affected by brain-gut (previously termed functional) disorders, including irritable bowel syndrome (IBS) and nonulcerative dyspepsia.6-9 In addition, several low-risk GI conditions, such as gastroesophageal reflux disease (GERD) and Helicobacter pylori infection, are commonly seen in primary care settings and patients with these conditions are often referred to specialty care.

However, many patients with common GI symptoms do not require subspecialty consultation or endoscopic investigation. Endoscopy is resource intensive and its limited availability in publicly funded health systems can be a key contributor to access bottlenecks. Furthermore, endoscopy findings are usually normal in patients with brain-gut disorders and contribute little to management.10 Several international guidelines recommend against endoscopic investigation of otherwise healthy individuals with chronic GI symptoms that lack alarm features.11-13 Instead, these individuals may be managed within the primary care medical home using appropriate support strategies such as evidence-based primary care clinical pathways (PCCPs).14,15

In Calgary, Alta, most GI and hepatology referrals are processed by Central Access and Triage (CAT), a single-point-of-entry system designed to manage referrals using standardized triage algorithms for accessing specialty care. In addition, carefully designed PCCPs that address low-risk GI referrals have been incorporated within CAT and aim to promote timely evidence-based care within the primary medical home. As little is known regarding the usefulness of GI PCCPs, the objective of this study was to evaluate the outcomes and health care use of patients whose referrals were declined by CAT and returned to their primary care physicians for management using PCCPs.

METHODSStudy design and setting

This study was a retrospective analysis of a patient cohort in Calgary. Calgary is the most populated urban centre in the Calgary Zone of Alberta Health Services, which is the largest integrated provincial health care system in Canada. The study was approved by the University of Calgary Conjoint Health Research Ethics Board (REB21-0601).

Overview of CAT and PCCPs

Central Access and Triage is a single-point-of-entry referral model encompassing 74% of all practising academic and community-based gastroenterologists (37 of a total of 50 physicians) in Calgary (personal communication with Greg Heather, Central Access and Triage Lead, Alberta Health Services, January 2022). Referrals are initially processed by trained registered nurses using triage algorithms who are overseen by a rotating roster of participating gastroenterologists. Results of standard, mandatory laboratory investigations are required for all referrals before protocolized assessment. These investigations typically include a complete blood count and, depending on the nature of the referral indication, often include additional investigations (eg, celiac disease screening, iron level, serum ferritin level, H pylori testing, C-reactive protein level, liver enzyme levels). Referrals for patients with important laboratory test abnormalities (eg, new iron deficiency anemia) and those with concerning clinical features such as weight loss, rectal bleeding, or family history of inflammatory bowel disease (IBD) or colorectal cancer are automatically accepted and prioritized as urgent. In contrast, patients without any concerning history and lacking “red flag” signs or symptoms, and who appear appropriate for a PCCP, are returned to the referring physician for management within the medical home. Patients referred without attachment to a primary care provider are accepted for assessment.

Since 2015 a total of 7 PCCPs have been co-designed by primary care providers and the Division of Gastroenterology and Hepatology at the University of Calgary to support current best evidence care for nonurgent routine referrals sent to CAT. The 7 PCCPs for the most common referral indications to general GI are chronic abdominal pain,16 chronic constipation,17 chronic diarrhea,18 dyspepsia,19 GERD,20H pylori infection,21 and IBS.22 Referrals deemed appropriate for a PCCP are initially declined (“closed”). Primary care providers are supported by access to same-day electronic advice (via telephone or secure email) from GI specialists 5 days a week. Patients who remain symptomatic despite management recommendations outlined in the pathway or who develop alarm features are accepted for streamlined consultation by a gastroenterologist.

Study participants

Between July 1, 2018, and May 31, 2020, consecutive adult patients (older than 18 years) referred by primary care providers whose referrals were initially closed and assigned to 1 of the 7 PCCPs were included. Each study participant had a valid personal health number matched to the Alberta Health registry. Patients were followed for a minimum of 30 days and up to 1 year from the referral closure date. For patients with more than 1 referral closure, the date of the last closed referral was used to determine the follow-up period.

Data collection and sources

Health care use and overall costs, including all emergency department (ED) and urgent care (UC) visits, hospitalizations, and endoscopic investigations, were calculated. Health care costs were calculated using the Canadian Institute for Health Information costing methodology, which employs costs per standard hospital stay and resource intensity weights.23

Data were extracted from multiple databases and linked via patients’ personal health numbers: Calgary CAT, the Alberta Health registry, the National Ambulatory Care Reporting System, the Discharge Abstract Database, physician billing claims, the EndoPRO endoscopy reporting system, and Sunrise Clinical Manager.

The closed referrals were abstracted from the CAT system, where the applied primary care pathway for each closed referral was captured, including the date the referral was closed and the referral indication. Patient demographic information was obtained from the Alberta Health registry, including sex, date of birth, and patient residence postal codes, which were mapped to a material and social deprivation index for socioeconomic status measures.24

Health care use was tracked following the closed referral date through examination of several administrative databases. Diagnosis codes from ICD-10-CA (International Statistical Classification of Diseases and Related Health Problems, 10th Revision, Canada) for the most responsible diagnosis associated with the captured health care events were reviewed (S.E.J.) to identify GI-related health care visits and admissions.

The EndoPRO endoscopy database was used to obtain endoscopy reports outlining procedure date, hospital site, indication, and endoscopist impressions. A few endoscopic procedures were completed outside the Calgary Zone (ie, were not reported using EndoPRO); as these results were not accessible, they were not included in the final analyses. The Sunrise Clinical Manager acute and ambulatory electronic health record was used to obtain pathology results. Endoscopy findings and associated histopathology reports were evaluated by 2 gastroenterologists (K.N., P.J.B.) to determine the presence of clinically relevant findings. When there was disagreement among reviewers the case was reviewed by a third gastroenterologist (S.V.V.Z.). Clinical significance was pre-defined as a diagnosis that would influence medical management, including those with immediate consequences (eg, malignancy) or long-term management implications (eg, IBD). Findings were deemed not clinically significant if they could have been diagnosed non-invasively, such as H pylori gastritis.

Statistical analysis

Descriptive statistics were reported as either proportions for categorical variables or medians with interquartile ranges for continuous variables. Health care costs incurred were accumulated through billing claims for visits, assessments, and procedures. The average direct health care cost of endoscopy, excluding physician fees, was calculated as $828. The cost of a gastroenterology consultation was $186.95; physician fees for gastroscopy and colonoscopy were $113.99 and $176.75, respectively. Costs avoided through closure of referrals were estimated based on the cumulative cost of a gastroenterology consultation and the assumption that 80% of these patients would have received endoscopic evaluation had their referrals been accepted. Thus, the cost avoided for each referral assigned to a PCCP was $113.99 + $176.75 + $186.95 = $477.69. Logistic regression was used to determine factors associated with ED and UC visits, hospitalizations, and performed endoscopies by examining age, sex, socioeconomic status, and initial GI referral indication.

RESULTS

During the study period, 3435 routine referrals were closed for 3274 patients. The cohort had a median age of 45 years; 46.4% were between the ages of 30 and 49 years. Most patients were female (66.1%), and 43.0% of patients were of higher socioeconomic status based on the material component of the deprivation index. Most patients (77.5%) resided within the Calgary urban core (Table 1).24

Table 1.

Patient demographic characteristics: N=3274; median age 45 y.

Table 2 describes the primary care pathways applied to the closed referrals. In some cases the referral indication was appropriate for more than 1 pathway, and in these instances information for each of the relevant PCCPs was sent to the referring primary care provider for a single referral. The most common closed referral indication was dyspepsia, applied to 1154 referrals, either alone or in combination with another pathway. Most often, the dyspepsia pathway was applied in combination with the GERD (80 of 159) or IBS (41 of 159) pathways.

Table 2.

Primary care pathways applied to closed routine referrals: A total of 3435 referrals were closed. Some closed referrals had >1 clinical care pathway attached to the closed or declined referral.

Health care use

Although follow-up periods varied, 1762 (53.8%) patients had 1 full year of follow-up. There were 362 patients (11.1%) who had 616 GI-related ED or UC visits. A total of 52 patients (1.6%) had 68 GI-related hospitalizations and 379 endoscopies were performed after re-referral (Table 3). Of the patient data available for analysis, most patients underwent endoscopies performed by a small number of community-based GI physicians or surgeons outside CAT (234 of 348; 67.2%).

Table 3.

GI-related health care use by patients within 1-y follow-up of closed routine CAT referrals: Individual patients may be represented more than once across columns and in >1 row.

Clinical outcomes

Of the 348 sets of endoscopic procedure findings available for analysis, 68.1% were normal without any endoscopic or histologic findings. Most incidental findings unrelated to the primary concern were low-risk colonic polyps (21.0%, 73 of 348). Relatively few (7.8%, 27 of 348) had clinically significant findings (Table 4). Three patients were newly diagnosed with colorectal adenocarcinoma. In these cases, the initial closed referral indication was completely different from the re-referral indication. One had a closed referral for dyspepsia, then was re-referred for a positive fecal immunohistochemical test (FIT) result, which was prioritized and appropriately investigated. The second closed referral was for chronic diarrhea, with a re-referral for new-onset hematochezia. The third closed referral was for dyspepsia, and the patient was re-referred for positive FIT result and rectal bleeding. Finally, a gastric large B-cell lymphoma was identified in a patient who had a referral closed the previous year for long-standing constipation.

Table 4.

Endoscopic procedure findings: N=348 patients had data available for analysis. Endoscopic outcomes available through EndoPRO Calgary Zone endoscopy database only.

Health care costs

The total health care cost avoided by closure of the 3435 routine GI referrals was estimated at $2,917,517. After up to 1 year of follow-up, the total health care cost accrued from the GI-related ED and acute care visits, hospitalizations, specialty re-referrals, and endoscopic evaluations for 362 patients of the 3274 patient cohort was $1,440,281 (Table 3). Therefore, the estimated health care cost avoided over the study period was $1,477,236. A sensitivity analysis challenging our assumptions that 100% would have received a consultation and 80% would have received endoscopy on referral showed a cost avoidance of $463,853 if only 75% of the referred cohort received a consultation and 50% received endoscopy.

Factors contributing to health care use

A total of 117 patients (3.6%) demonstrated higher use within 1 year of their referral being closed, defined as 2 or more visits to the ED or UC. In adjusted multivariable analysis, younger age (P=.01) and place of residence (P=.03) were associated with higher ED and UC use. Those residing in the South Zone of Alberta (a predominantly rural area) had greater odds (odds ratio [OR] 3.71, 95% CI 1.61 to 8.53) of repeat ED and UC visitation compared with patients residing in the Calgary core (Table 5).24

Table 5.

Logistic regression assessing patient factors associated with patients making 2 or more ED or UC visits

A second logistic regression evaluating patient factors driving increased likelihood of re-referral with endoscopic examination was undertaken (Table 6).24 There were 348 (10.6%) patients who received at least 1 endoscopy during the 1-year follow-up period. Age and applied pathway were statistically significant factors (P=.032 and P=.004, respectively). Patients aged 30 to 49 were less likely to receive an endoscopy compared with patients aged 70 and older, with an OR of 0.70 (95% CI 0.50 to 0.98). Patients who had either dyspepsia or GERD pathways applied to their referral were associated with greater odds of undergoing endoscopy compared with patients with referrals closed for the IBS pathway, with ORs of 1.60 (95% CI 1.11 to 2.32) and 2.07 (95% CI 1.42 to 3.03), respectively.

Table 6.

Logistic regression assessing patient factors associated with patients receiving ≥1 endoscopies

DISCUSSION

As demonstrated here, many patients referred from primary care with common, chronic GI symptoms can be safely managed by primary care providers with guidance from clinical care pathways. As expected, not all remained within their medical homes: Of the 3274 patients managed with a PCCP, 21.7% presented to an acute care facility, were seen in consultation, or underwent endoscopy during up to 1 year of follow-up. Among these, a low rate of patients presented with important clinical findings, as 9 patients (0.3% of the total cohort) were diagnosed with a serious condition, including 4 cancers and 4 cases of IBD. Thus, with appropriate consistent communication and collaboration between primary and specialty GI care, patients who would otherwise wait in long queues may be managed with pathways reflecting best-evidence care.

The portion of the patient cohort (11.1% of 3274) presenting to acute care may suggest high use of health care services. It is possible patients seek ED or UC services for their symptoms in the absence of guidance from either specialty or primary care. For this cohort, we have no detailed data regarding the presenting concern or reason for health system use. Therefore, the health system use accounted for here may be unrelated to their original closed referrals, resulting in higher perceived use rates. Avoidable health system use is a common challenge for both specialty and primary care; therefore, collaborative efforts must be undertaken to jointly support patients and avoid presentation for chronic concerns.

Patients re-referred for endoscopy had completed the pathway with persistent symptoms, had progression of symptoms, or were referred for an alternate indication such as cancer screening. Four cancer cases and 4 new diagnoses of IBD were identified among the 10.6% of patients who were re-referred. Importantly, nearly all diagnosed cases were unrelated to patients’ original referrals. Each referral for malignancy was unrelated to the original closed referral. These cases each exhibited new-onset alarm symptoms (eg, new hematochezia) or positive FIT results, warranting endoscopy. Three of the 4 patients with new IBD diagnoses warranted urgent investigation due to bleeding or weight loss, while the fourth patient reported new diarrhea at the age of 79 years. These patients’ cases highlight the essential need for consistent follow-up within primary care, as symptoms can change and evolve, requiring attention.

On logistic regression analysis, we found higher ED and UC use among younger patients (aged 16 to 29 years) and those living in rural areas. It is possible patients may have more limited access to primary care in rural areas, resulting in greater use of UC or the ED. Further studies to better understand drivers among younger patients living in rural regions are needed.

Patients assigned to the dyspepsia and GERD pathways were more likely to be re-referred for endoscopic evaluation. This may relate to the fact that apart from acid suppressive therapy and lifestyle changes, which can be difficult to achieve (eg, weight loss), relatively few evidence-based therapies exist. Furthermore, long-standing GERD symptoms have historically been considered an indication for gastroscopy, despite recent guidelines limiting gastroscopy when risk factors for metaplasia are present, such as male sex, obesity, and age 50 years and older.25,26 In patients younger than 50 years at low risk, gastroscopy should be avoided.25,26 Our study provides support for the use of PCCPs as a strategy to facilitate evidence-based management of patients with low-risk GI conditions such as GERD. Importantly, our analysis also points to substantial health expenditures being avoided through fewer GI consultations and reductions in low-yield procedures.

System-level innovations are urgently needed to address expanding gaps in access and longer wait times for specialist care. Historically, increased demand for GI specialty consultation and investigation has been addressed through advocacy to expand resources and through efforts to increase efficiency (eg, bringing patients directly to endoscopy with consultation occurring during the same appointment). There are limited resources available to support Canada’s publicly funded system and, to date, expansion of services has had minimal success in reducing wait times.2 Moreover, the COVID-19 pandemic led to further increases in wait times across Canada, particularly for those living in rural locations.27 For many jurisdictions it simply may not be possible to address the backlog. Now more than ever our health system must prioritize high-quality care, limit the use of low-yield investigations, and provide primary care physicians with timely resources and support to manage patients within the medical home.

The importance of patient education, reassurance, and validation by primary care physicians cannot be overemphasized.28,29 Clear communication regarding the need for referral and the referral process itself, in addition to validating a patient’s symptoms and providing reassurance when appropriate, are all crucial patient engagement components. Primary care providers are the most important puzzle piece, and without them the use of PCCPs will not improve access to care. Many symptomatic patients worry about the potential for serious underlying disease and thus believe they need to undergo comprehensive evaluation by a consultant, including endoscopy. Further, patients with closed referrals may perceive that they have not been heard or that their concerns are unimportant, as opposed to simply receiving confirmation of lower-risk status. Support and direction from primary care providers is essential to outline the lack of need or contribution of endoscopic examination as it adds little to understanding or treatment of symptoms while exposing patients to unnecessary risk. Finally, it must be stressed that a closed referral does not preclude re-referral should symptoms change, additional concerning findings arise, or there is a clear nonresponse to instituted therapies. Primary care providers must feel supported, and continued communication and collaboration are essential for high-quality patient care. Given the importance of collaboration in determining the long-term success of PCCPs, we are exploring patient and physician perceptions regarding closed referrals and other possible drivers of ED and UC use in greater depth.

Limitations and strengths

There are several limitations to this study. First, the available data are aggregate and administrative, and a deeper understanding of individual characteristics (eg, treatments attempted) through manual chart reviews was not possible. Confirmation of a direct linkage between the individual pathway indication to the ED presentation or reason for hospitalization was also not possible. In addition, our clinical follow-up was relatively short and thus additional patients may have presented to the ED or have undergone endoscopy had follow-up been lengthened. However, it is likely that most serious diagnoses such as cancer would have manifested during our study time frame. Finally, neither primary care provider nor patient experience were evaluated in this study, although qualitative evaluation of this cohort is under way. Results from such work will inform further improvements to pathway implementation, communication to patients and primary care providers, and the overall process. Despite these limitations, there are several strengths. This study details a retrospective analysis of a large, real-world experience in an urban centre using a well-established and broadly deployed centralized triage system within an engaged and participatory primary care community. This is the first large-scale study evaluating this innovative model in Canada.

Conclusion

Collaboratively developed PCCPs can be implemented safely for patients with routine, low-risk GI indications and carry the potential for avoidance of substantial health care costs. Widespread adoption of PCCPs for patients with low-risk GI referrals has the potential to optimize waitlist management and reduce low-yield endoscopy to enable considerable cost savings and resource redeployment to use the limited resource of GI specialist care more judiciously.

Footnotes

Contributors

All authors made substantial contributions to the concept and drafting of the manuscript. All authors approved the final version for submission.

Competing interests

None declared

This article has been peer reviewed.

Cet article a fait l’objet d’une révision par des pairs.

Copyright © 2024 the College of Family Physicians of Canada