Article Text
Abstract
Background Events occurring outside the hospital setting are underevaluated in surgical quality improvement initiatives and research.
Objective To quantify regional variation in home care nursing following vascular surgery and explore its impact on emergency department (ED) visits and hospital readmission.
Methods Patients who underwent elective vascular surgery and were discharged directly home were identified from population-based administrative databases for the province of Ontario, Canada, 2006–2015. The index surgeries included carotid endarterectomy, open and endovascular aortic aneurysm repair and bypass for lower extremity peripheral arterial disease. Home care nursing within 30 days of discharge was captured and compared across regions. Using multilevel logistic regression, we characterised the association between home care nursing and the risk of an ED visit or hospital readmission within 30 days of discharge.
Results The cohort included 23 617 patients, of whom 9002 (38%) received home care nursing within 30 days of discharge home. Receipt of nursing care after discharge home varied widely across Ontario’s 14 administrative health regions (range 16%–84%), even after accounting for differences in patient case mix. A lower likelihood of an ED visit or hospital readmission within 30 days of discharge was observed among patients who received home care nursing following three of four index surgeries: carotid endarterectomy OR 0.74, 95% CI 0.61 to 0.91; endovascular aortic aneurysm repair OR 0.85, 95% CI 0.72 to 0.99; open aortic aneurysm repair OR 1.06, 95% CI 0.91 to 1.23; bypass for lower extremity peripheral arterial disease OR 0.81, 95% CI 0.72 to 0.92.
Conclusion Home care nursing may contribute to reducing ED visits and hospital readmission and is variably prescribed after vascular surgery.
- surgery
- quality improvement
- nurses
- healthcare quality improvement
Statistics from Altmetric.com
Introduction
Hospital readmission after surgery is a prominent focus of quality improvement initiatives and research.1–3 However, events occurring outside the hospital setting remain underevaluated. Efforts to improve outcomes for surgical patients would be strengthened by shifting focus beyond hospital care towards the full trajectory of a patient’s recovery, including at-home or in-community health needs. In doing so, a fully integrated care pathway can be held to account. Furthermore, as public healthcare systems and third-party insurance providers seek to reduce hospital length of stay, recovery following surgery increasingly occurs at the patient’s home rather than in the hospital.4 5 However, the indications and impact of even the most common postoperative outpatient services, such as home care nursing care, remain underevaluated.
Given these considerations, we sought to study home care nursing following elective vascular surgery within a single-payer public healthcare system. We hypothesised that home care nursing after vascular surgery is variably prescribed but may reduce emergency department (ED) visits and hospital readmission.
Methods
Study design and setting
We designed a population-based retrospective cohort study of patients who were discharged directly home following elective vascular surgery in the province of Ontario, Canada, between 1 April 2006 and 31 March 2015. The index surgeries were four common interventions with differing needs for nursing care after discharge—carotid endarterectomy, endovascular aortic aneurysm repair (EVAR), open abdominal aortic aneurysm (AAA) repair and bypass for lower extremity peripheral arterial disease (PAD). Home care nursing, delivered at the patient’s home or in their community, was captured within 30 days of discharge home, as were ED visits and hospital readmission.
Ontario is Canada’s most populous province with over 13 million residents living across over 1 million km2, an area larger than France and Spain combined. All hospitalisations for vascular surgery and outpatient nursing services following surgery are publicly funded under the province’s single-payer healthcare system. Oversight and delivery of outpatient home care services are structured within 14 administrative health regions, known as Local Health Integrated Networks.
Variation in receipt of home care nursing was first evaluated across these administrative health regions. We then explored the association between receipt of home care nursing and the risk of an ED visit or hospital readmission within 30 days of discharge.
Data sources
The index hospitalisation for vascular surgery was identified from the Canadian Institute for Health Information’s Discharge Abstract Database. The Discharge Abstract Database includes a unique record for every hospital admission in Ontario with data on a patient’s diagnoses (most responsible, comorbidities, those arising in-hospital), procedures and discharge disposition. Patient demographics (age, sex, rurality of residence, neighbourhood marginalisation index) were obtained from the Registered Persons Database. The Ontario Health Insurance Plan Physician Claims database, which includes fee claims for all physician services in Ontario, and the Discharge Abstract Database contributed to characterising comorbidity level based on inpatient and outpatient records. All nursing visits occurring in-home or in-community after discharge were available from the Home Care Database. ED visits were captured in the National Ambulatory Care Reporting System Database, which contains a unique record for every ED visit in Ontario. There are no private hospitals in Ontario offering ED or inpatient vascular services.
The data sets were linked using unique, encoded identifiers and analysed at ICES.6 Deterministic linkage of these data sets has been validated for the identification of a variety of individual diagnoses and surgical procedures including the index vascular surgeries—positive predictive values (PPV): 99% for carotid endarterectomy, 95% PPV for open AAA repair, 96% for EVAR, sensitivity and 75%–88% for bypass.7–9However, the accuracy of home care nursing visit capture in the Home Care Database has not been validated.
Cohort
The cohort is based on a previously described elective vascular surgery cohort among adult Ontario residents.10 The index hospitalisation for surgery was required to include specified most responsible diagnosis and procedure codes for: carotid endarterectomy for cerebrovascular atherosclerotic disease, EVAR for non-ruptured AAA, open AAA repair for non-ruptured AAA and bypass for PAD.10 Exclusions included: (1) surgery to the same arterial bed within a 2-year look-back period, since reoperative patients may reflect previous technical complication and be outliers with respect to risk of an ED visit or hospital readmission; (2) hospitalisations over 90 days, since these patients are also atypical outliers; (3) death on index hospital admission; and (4) less than 30 days between hospital discharge and maximum available follow-up date. The International Classification of Diseases 10th Revision diagnosis codes and Canadian Classification of Health Interventions codes defining these inclusion and exclusions are previously published.10 Additional exclusions required in order to examine nursing services after discharge home included: (1) hospitalisation prior to 1 April 2006, the date of inception of the Home Care Database, and (2) discharge disposition other than home.
Home care nursing within 30 days of discharge
All home care nursing delivered anywhere in the province of Ontario within 30 days of hospital discharge was considered. Home care nursing, more appropriately termed in-home or in-community nursing care, included nursing care for wounds, intravenous injection of antibiotics or other medications as well as, less commonly, some chronic disease teaching (eg, diabetes, peritoneal dialysis). Prescription of these services can be initiated by the treating surgeon, a nurse practitioner, family physician or any other physician involved in the patient’s care (eg, emergency room physician). There are no standard indications for a home care nursing prescription after surgery and there are no restricted criteria for funding across Ontario’s administrative health regions. Communication between home care nurses and the prescribing physician is common and usually completed in writing or by phone. Nursing visits can be stopped at the discretion of the patient or nursing team but a defined physician follow-up date is always required.
ED visit or hospital readmission within 30 days of discharge
All ED visits or acute care hospitalisations, at the hospital of the index surgery or any of other hospital in the province of Ontario, were captured within 30 days of index hospital discharge.
Covariates
A number of covariates were captured in order to account for differences in patient case mix between regions as well as to adjust for potential confounders of the association between home care nursing and the risk of an ED visit or hospital readmission.
The following covariates were defined: patient demographics (age, sex, rurality of residence, Ontario Marginalization Index), patient comorbidity level (Aggregate Diagnosis Group Comorbidity Index, home care nursing within 30 days before the index hospitalisation for surgery) and measures related to complexity of the index hospitalisation (teaching hospital, hospital length of stay, intensive care unit length of stay more than 1 day, weekend discharge). Fiscal year was also captured given potential for changes over time in nursing care prescription patterns. All covariates were defined relative to the date of the index hospitalisation.
The Statistics Canada dissemination area, a geographic measure much smaller than the administrative health region and encompassing 400–700 residents, was the basis for information on rurality of residence and the components of the Ontario Marginalization Index (dependency, material deprivation, ethnicity, residential instability).11 12 These measures may reflect potential differences in a patient’s access and proclivity to access healthcare resources. The Ontario Marginalization Index has been associated with behaviours that impact on health (eg, smoking, obesity, receiving the annual influenza shot) and health outcomes.12 The John Hopkins Adjusted Case-Mix System, applied to inpatient and outpatients records within the 2 years prior to index hospitalisation, was used to calculate a comorbidity index validated for prediction of 1-year mortality in adult Ontarians.13 14 In addition to hospital teaching status, defined as academic teaching or non-teaching by the Ontario Ministry of Health,15 hospital length of stay, intensive care unit admission and weekend discharge were considered crude surrogates for complexity of the patient’s inpatient recovery.
Analysis
We first characterised the crude frequency of any home care nursing within 30 days of discharge for the full cohort, each of the four operation subgroups and across the 14 administrative health regions. Variation across regions was represented in funnel plots where each region’s proportion of patients receiving home care nursing was plotted against the region’s volume of index vascular surgeries. Ninety-nine per cent control limits frame the range of random variation around the overall frequency of nursing care for the entire cohort.16 Variation in receipt of nursing care across regions, adjusted for expected differences in patient case mix between regions and temporal changes, was further characterised by calculating the median OR from a hierarchical (two-level) multivariable logistic regression model. The binary outcome was receipt of nursing care within 30 days of discharge home, with a random intercept at the region level.17 18 The previously listed patient and hospitalisation covariates as well as surgery type and fiscal year were all selected a priori for model inclusion given their potential influence on receiving home care nursing. Goodness of fit was evaluated with an observed-versus-expected outcome plot and showed good calibration. The median OR is the median value of all possible ratios of the odds of receiving nursing care for two patients, with the same covariate values, treated in two randomly selected distinct regions. By convention, the median OR is always greater than or equal to 1 since the odds of the patient in the region with the greatest proclivity for nursing care are used as the numerator. As an example, a median OR of 2.0 suggests a twofold median difference in the odds of receiving nursing care for two similar patients treated in distinct randomly selected regions.
Next we explored the potential impact of home care nursing on the likelihood of an ED visit or hospital readmission within 30 days of discharge. In this analysis, any nursing care delivered after the first ED visit or hospital readmission was excluded from the exposure definition. A second hierarchical (two-level) logistic regression model was specified to characterise the association between receipt of home care nursing (exposure) and the risk of an ED visit or hospital readmission within 30 days of discharge (outcome), while accounting for clustering of similar patients within a given region (random intercept). Given expected differences in the strength of association between the exposure and outcome across the four surgery types, surgery type was treated as an effect modifier (interaction term with exposure) in the model. The model was specified to include all potential confounders of the exposure-outcome relationship. All previously listed covariates were retained as potential confounders based on a priori clinical grounds. Goodness of fit was evaluated with an observed-versus-expected outcome plot and showed good calibration.
All statistical tests were two sided, with statistical significance defined as p value <0.05. Analyses were performed using SAS V.9.4 (SAS Institute).
Results
Cohort
A total of 23 617 patients met the inclusion and exclusion criteria (figure 1). This cohort included 7223 carotid endarterectomies, 4861 EVARs, 5008 open AAA repairs and 6558 lower extremity bypasses.
Home care nursing within 30 days of discharge home
Within 30 days of hospital discharge, 9002 (38% of 23 617) patients received at-home or in-community nursing care. As expected, this proportion differed across surgery types: 68% (n=4461 of 6558) in lower extremity bypass, 36% (n=1813 of 5008) in open AAA repair, 33% (n=1588 of 4861) in EVAR and 16% (n=1161 of 7223) in carotid endarterectomy.
Regional variation in home care nursing within 30 days of discharge home
Receipt of home care nursing after discharge varied widely across administrative health regions (figure 2). Baseline differences between patients receiving nursing care existed based on surgery type, patient characteristics and measures related to duration and complexity of the hospitalisation (table 1). After accounting for these differences in case mix and year across regions through multivariable logistic regression (online supplementary table 1), residual variation remained as evidenced by a median OR of 2.5. In other words, a similar patient treated in one randomly selected region versus another had a 2.5-fold median difference in the odds of receiving nursing care after discharge.
Supplemental material
The median number of nursing care days was 6 (IQR 4–12, maximum 30). There was no consistent correlation between the intensity of nursing care (ie, number of nursing visit days out of 30) and a health region’s proportion of patients receiving home care nursing (table 2).
Home care nursing and the risk of an ED visit or hospital readmission within 30 days of discharge home
At a regional level, the lowest risk of an ED visit or hospital readmission within 30 days was observed in regions with the greatest proportion of patients receiving nursing care after discharge (table 2).
At the patient level, home care nursing and the risk of an ED visit or hospital readmission varied across surgery types (table 3). Multivariable analysis accounted for overall negative confounding (ie, bias towards the null) and demonstrated that home care nursing was associated with lower likelihood of an ED visit or hospital readmission within 30 days of discharge in three of four index surgeries: carotid endarterectomy, EVAR and bypass for lower extremity PAD (table 3 and online supplementary table 2).
Discussion
This population-based study focused on home care nursing (at-home or in-community) after discharge as a quality-of-care measure following vascular surgery in Ontario, Canada. There were two main findings. First, there was a significant variation in receipt of nursing across administrative health regions, even after accounting for regional differences in patient case mix. Second, home care nursing was associated with a lower risk of an ED visit or hospital readmission within 30 days of discharge for carotid endarterectomy, EVAR and bypass for lower extremity PAD but not open aortic aneurysm repair.
The observed association between home care nursing and a reduced risk of an ED visit or hospital readmission represents a unique opportunity to better support vascular surgery patients as they recover at home. A few others have looked at medical patients receiving home care nursing and identified either no impact, or an increase, in the risk of an ED visit or hospital readmission.19–21 However, surgical patients may differ from medical patients in two important ways. First, home care nurses may be able to more effectively address early postdischarge issues in surgical compared with medical patients (eg, bleeding from wound, concern about surgical site infection). Second, surgical patients usually have one most responsible physician, the surgeon, who provides both inpatient care and outpatient follow-up. However, medical patients often have outpatient follow-up provided by their family physician or a specialist who differs from their inpatient provider. The continuity of care across inpatient and outpatient settings in surgical patients may facilitate clear communication of a home care nursing treatment plan thereby avoiding ED visits and readmission. It should be noted however that our study demonstrates an inconsistent benefit from home care nursing across different types of surgery. In fact, Sanford et al corroborate this finding with their study showing that home health services increased the risk of readmission following pancreatectomy among Medicare patients.22 The impact of home care nursing should be expected to differ across surgery types because the causes and frequency of readmission differ across surgeries. While these hypotheses cannot be proven based on our data, it seems reasonable to suggest that further research on the potential benefit of home care nursing should be undertaken across a broader range of surgical procedures.
There are also specific implications from this analysis for vascular surgical care in Ontario. It behoves the vascular surgical community and healthcare system administrators to better understand outlier regions where, for example, over 75% of patients receive home care nursing compared with less than 20% of patients in other regions. Overprescription of nursing care can burden patients with home visits or travel to wound care clinics, reduces the availability of home care nurses for other patients and inflates the costs of care. At the same time, the association of nursing care and lower ED visit or hospital readmissions suggests a potential benefit from home care nursing. More work is required to understand the underlying reasons for this benefit in order to standardise the indications for nursing after vascular surgery across all regions in Ontario.
The study findings also underscore the fact that home care nursing merits focus as a process measure of quality of care. In the era of big data, efforts to improve the quality of surgical care increasingly rely on administrative or registry data capturing both process and outcome measures. For example, there are numerous publications characterising variation and risk factors for hospital readmission following surgery in the National Health Service of the UK, or under Medicare in the USA. In contrast, almost no peer-reviewed studies discuss the home care services received following surgery for these same patients despite the existence of funding for home health services under both the National Health Service and Medicare.23 24 The aforementioned pancreatectomy study and the Michigan Urological Surgery Improvement Collaborative are the two notable exceptions.22 25 A relative lack of emphasis on outpatient care in surgical quality improvement may be due to a number of factors, which are often context specific. On one hand, a hospital-based surgical team may have a limited ability to track outcomes occurring outside the walls of their hospital. In other instances of multi-institution external benchmarking, the hospital perspective dominates since funding support for the surgical quality improvement programme is provided by the hospital.26 Our analysis demonstrates the useful insights that can be gained when considering both acute care as well as at-home or in-community health needs in surgical quality improvement.
This evaluation of home care nursing after vascular surgery was strengthened by the scope of available data within our publicly funded regional healthcare system. However, certain important limitations warrant mention. First, the indication for a nursing visit was not available from the Home Care Database. As a result, the current analysis cannot, for instance, clarify whether the indication for nursing care was to prevent or treat a complication, nor do we know what exact care a nurse delivered. Second, despite the range of validated covariates included in our multivariable model, residual confounding due to unmeasured characteristics may partly explain the observed association between nursing case and the risk of an ED visit or hospital readmission. Unmeasured potential confounders include preoperative disease severity (eg, Wound, Ischemia, Foot Infection score in patients with PAD), functional limitation at discharge and specific wound and infectious complications during the index hospitalisation. Nevertheless, we have attempted to account for these confounders through surrogate variables including preoperative nursing care and the index admission characteristics. Third, additional home nursing services can be paid for privately and supplement what is funded by the public healthcare system. These private services would not be captured in the Home Care Database. However, nursing care for wounds including necessary supplies requires a physician prescription and is fully funded across all regions of Ontario so private nursing care is likely uncommon.
In conclusion, within a single-payer regional healthcare system, similar patients undergoing vascular surgery inconsistently received nursing care after discharge home. Those who did receive home care nursing were less likely to present to an ED or be readmitted to hospital within 30 days of discharge following carotid endarterectomy, EVAR and bypass for PAD. These novel results represent an important opportunity to improve outpatient recovery following vascular surgery.
Acknowledgments
This study made use of de-identified data from the ICES Data Repository, which is managed by the ICES with support from its funders and partners: Canada’s Strategy for Patient-Oriented Research (SPOR), the Ontario SPOR Support Unit, the Canadian Institutes of Health Research and the Government of Ontario. The opinions, results and conclusions reported are those of the authors. No endorsement by ICES or any of its funders or partners is intended or should be inferred. Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information (CIHI). However, the analyses, conclusions, opinions and statements expressed herein are those of the author, and not necessarily those of CIHI.
References
Footnotes
Contributors CdM: design of the work; acquisition, analysis and interpretation of data; drafted the work. GRN: design of the work; acquisition and interpretation of data. AK, MAO: design of the work and interpretation of data. KS, MAH: interpretation of data. CdM, AK, MAO, KS, MAH, GRN: revised the work critically for important intellectual content. CdM, AK, MAO, KS, MAH, GRN: final approval of the version published. All authors agree to be accountable for all aspects of the work.
Funding The study was supported by the Blair Foundation Vascular Surgery Innovation Fund jointly established with the University of Toronto
Competing interests None declared.
Patient consent for publication Not required.
Ethics approval ICES is a prescribed entity under section 45 of the Personal Health Information Protection Act of Ontario. Under section 45, ICES is allowed to collect and use personal health information from health information custodians for the purposes of evaluation, planning and monitoring of the health system. The approval of a research ethics board was therefore not legally required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data may be obtained from a third party and are not publicly available.