Economic Analysis of Home vs Hospital-Based Parenteral Nutrition in Ontario, Canada 
From the * Department of Medicine, Division of
Gastroenterology, McMaster University; the Correspondence: Dr. John K. Marshall, Division of Gastroenterology (4W8), McMaster University Medical Centre, 1200 Main St. W, Hamilton, Ontario L8N 3Z5, Canada. Electronic mail may be sent to marshllj{at}mcmaster.ca. Background: Advances in technology and infrastructure have facilitated transfer of complex services from acute care hospitals to the home. This increases the burden on community resources but may provide net savings to the health care system. We undertook a retrospective cohort study of patients transferred from hospital to home while receiving home parenteral nutrition (PN) to assess their costs of care. Methods: A detailed review of medical records was undertaken for all patients managed by the Hamilton Health Sciences Home PN Program between 1996 and 2001 whose PN was initiated in hospital. Mean per diem direct medical costs were estimated from the perspective of the provincial Ministry of Health for 3 periods: the last 2 weeks before discharge and the first month after discharge. Costs were compared among time intervals and among patients subgroups defined by age and underlying disease. Results: Twenty-nine eligible subjects were identified. Common indications for PN included malignancy (n = 12), inflammatory bowel disease (n = 6), and intestinal ischemia (n = 4). Mean per diem costs in the last week of hospitalization were higher than those in the first month after discharge ($567 vs $405, p < .0001). Acute care resources accounted for <10% of the overall costs on home PN. The estimated monthly savings per patient maintained on home PN were $4860 (95% confidence interval $2700–$7000). Savings were even greater among patients with underlying malignancy and advanced age. Conclusions: Home PN is cost saving when compared with hospital-based PN. Neither age nor underlying malignancy should pose a barrier to receipt of home PN. Recent advances in technology and improvements in community support infrastructure have facilitated the transfer of many complex services from acute care hospitals to community providers. Parenteral nutrition (PN) can be used to sustain patients with intestinal failure for an indefinite period, but requires advanced multidisciplinary support. Many jurisdictions worldwide have succeeded in developing programs to support and supervise the delivery of PN in the home. Such programs are resource intense and are perceived by some to be prohibitively costly. However, they may still yield cost savings when compared with chronic hospitalization.1,2 Previous cost analyses have concluded that home PN reduces costs by 65%–81% when compared with hospital-based PN.2–7 However, these estimates differ in the components of cost they included, the assumptions they made, and in the analytic perspective they assumed. For example, Curtas et al estimated personnel costs of home PN delivery but not the overall costs of care.8 The few studies that assessed a comprehensive set of costs from the perspective of a third party2,4,6 are now considerably dated and may not reflect a current standard of PN care. We undertook a retrospective cohort study of patients transferred from hospital to the home while receiving PN in order to identify the costs associated with conversion to home PN delivery, from the perspective of a Canadian government provider.
All patients managed by the Hamilton Health Sciences (HHS) Home PN Program between 1996 and 2001 were identified from an existing program registry. Consecutive eligible subjects were required to have initiated PN in an acute care hospital before their discharge. The HHS Home PN Program comprises a nurse, a dietitian, and a physician, who work in collaboration with regional community care access centers and pharmacists to administer, supervise, and coordinate homecare services. The study protocol received full approval from the Research Ethics Board of Hamilton Health Sciences. Eligible patients provided written consent for review of their records and health information. In the case of patients who died after initiating home PN, consent for review of their records was obtained from the next of kin.
Data Collection
Inpatient Per Diem Costs Patient acuity was estimated for each day of inpatient care using the MEDICUS system (Quadra Med Corporation, San Rafael, CA), which combines 36 patient care parameters to derive an ordinal acuity score from 1 to 6. For each day of hospitalization, hospital hotel costs specific to each individual subject's acuity and ward location were obtained from the Decision Support Department of London Health Sciences Centre (LHSC). LHSC is a teaching hospital similar in size and acuity to HHS that uses a standardized, fully-allocated accounting system to maintain a database of case costs and individual resource unit costs indexed by individual patient encounter. LHSC also uses MEDICUS to weight patient acuity. Unit costs for procedures, tests, and laboratory and transfusion services were also obtained from LHSC. Costs per unit dose of medications and IV solutions were obtained from the HHS inpatient pharmacy with a 10% markup. Unit costs for billable physician services were obtained from the Ontario MOHLTC Schedule of Benefits (2001). Home PN nurse and dietitian costs were estimated using standard hourly wages with benefits.
Outpatient Per Diem Costs Costs for contracted services such as nursing, occupational therapy, physiotherapy, nutrition, and homemaking were estimated from patient-specific home care invoices. An overhead rate of 10% was added to each contract for CCAC administration. Costs for medications were obtained from the Ontario Drug Benefit formulary with standard dispensing fees ($6.11–$2.00 patient copayment). Unit costs for laboratory and transfusion services were obtained from the 2001 Ontario MOHLTC Schedule of Laboratory Fees. Costs for hospital readmissions were estimated by multiplying the length of stay by the mean patient-specific per diem cost estimated for week –1. Costs for other acute care services, physician services, and home PN nurse and dietitian services were described as for inpatient costs.
Analysis
A total of 29 eligible subjects were identified and included in the analysis. Among these, the mean age was 51 (range 18–81) years and 18 patients (62%) were female. Underlying diagnoses contributing to intestinal failure included malignancy (12 subjects; 41%), inflammatory bowel diseases (6 subjects; 21%), intestinal ischemia (4 subjects; 14%), fistula (3 subjects, 11%), and other diagnoses (4 subjects; 14%) (Table I). All subjects survived for at least 1 month after discharge.
Inpatient Costs
Subgroup analyses revealed that mean inpatient per diem costs for patients with malignancy were higher than those for patients with other indications for PN ($644 ± $186 vs $513 ± $85, p = .021). In patient per diem costs did not differ significantly by age or sex.
Outpatient Costs
In subgroup analyses, the mean per diem costs in month +1 for patients with intra-abdominal malignancy were higher than those with other underlying diagnoses ($463 ± $146 vs $364 ± $100, p = .038). Costs were also higher for subjects <55 years than for subjects age 55 or older ($452 ± $154 vs $355 ± $69, p < .04).
Comparison of Inpatient vs Outpatient Costs The mean per diem costs for month +1 were significantly lower than those for week –1 (paired analysis, p < .0001) (Fig. 1). Savings accrued through home PN were estimated on the assumption that, without hospital discharge, per diem hospital costs would continue at the rate observed in week –1. By this approach, the net savings per patient accrued during the first month after discharge were estimated to be $4860 (95% CI; $2700–$7000). These savings increased to $5400 per patient (95% CI; $360–$10400) among those with underlying malignancy and to $7170 per patient (95% CI; $4350–$9960) among those aged 55 or older.
The delivery of home PN has progressed considerably since its use was first reported over 30 years ago.9 The current study provides a rigorous, detailed, and up-to-date economic analysis of home PN delivery. By assessing a consecutive cohort of patients transferred from hospital to home while receiving PN, paired analyses could assess the cost impact while controlling for interpatient variability. Our results suggest that conversion to home PN accrues substantial savings to a provincial Ministry of Health, totaling approximately $4860 Canadian per patient over the first month after discharge. Hence, a home PN program with capacity for 30 clients could yield savings in excess of $1.7 million Canadian per year. Savings were greater among elderly subjects and among those with underlying cancer, suggesting that neither age nor malignancy should represent a barrier to the delivery of home PN from a cost perspective. This study has several shortcomings that warrant recognition. First, its retrospective design relies on adequate documentation of resource use in patient charts and electronic medical records. Every attempt was made to construct comprehensive resource profiles through meticulous and detailed microcosting, but some resource use could have been overlooked. Second, although the prepost design allowed intrapatient comparison and reduced error from interpatient variability, a prospective randomized trial comparing hospital vs home PN would have been more rigorous and would have avoided bias from temporal trends in health status and health resource requirements. We used week –1 resource profiles to estimate the cost of hospital PN under the assumption that health status immediately before discharge was otherwise stable, as this is a requirement for home PN access. Third, because we assessed costs from the perspective of a government payer, costs borne by nongovernment payers (eg, private insurers) and indirect costs associated with changes in productivity are not considered. Finally, the short time horizon (1 month after discharge) may overestimate the costs of long-term home PN. As patients and their caregivers become more familiar with managing PN, they could gain independence and require fewer community resources. These data originate from a single home PN program in Ontario, Canada. Because estimates of cost vary substantially among healthcare jurisdictions, and because many models exist for the funding and delivery of home PN, it should not be assumed that our favorable results can be applied elsewhere. Our conclusions are consistent with those of other analyses comparing home vs hospital-based PN3–7 and provide strong justification for stable funding of home PN services in Canada and elsewhere. The relatively small size of our study cohort provides limited statistical power for subgroup analysis. Although results should be interpreted with caution, we demonstrated greater savings among patients with advanced age and underlying malignancy. Outcomes on home PN are determined largely by progression of underlying disease, and less by demographic parameters like age.10 However, both advanced age and an underlying diagnosis of malignancy are sometimes considered contraindications to home PN, in part due to anticipated complications requiring acute care resources. In some jurisdictions, advanced malignancy has become an increasingly common indication for home PN,10 but this remains controversial. From an economic perspective, our results suggest that neither age nor malignancy should be a barrier to receipt of PN. Although this study demonstrates substantial cost savings with conversion from hospital to home PN, it did not directly assess change in other outcomes such as health status and patient/caregiver satisfaction. Howard has suggested that health outcomes on home PN in the context of a well-structured program are similar to those achieved with hospital-based PN.11 We observed little requirement for acute care resources and few complications directly attributable to PN. Although the health-related quality of life of patients who require home PN is worse than that of the general population, the limited published data suggest that it is superior to that of similar patients maintained on hospital-based PN.12–14 Hence, home PN may offer an alternative to hospital-based PN that is "dominant" in economic parlance, with both reduced costs and improved health outcomes.2,7 A prospective study of health outcomes and health-related quality of life is now underway at our center.
This study demonstrates that home PN is an economically favorable alternative to hospital-based PN, and accrues substantial cost savings within the first month after discharge. Despite the complexity of PN delivery and the substantial comorbidity of its recipients, consumption of acute care resources while on home PN is small and represents only a fraction of overall costs. Among patients who might be predicted to do poorly on home PN (eg, elderly patients and those with underlying malignancy), the overall costs of care are higher, but the net savings accrued by conversion to home PN are even more substantial. Improved funding and access to home PN services is justified. Financial support was provided by an operating grant from the Hamilton Health Sciences Foundation. The authors thank the Decision Support Department at London Health Sciences Centre for their assistance in obtaining resource unit costs, and Ms. Marroon Thabane for her help with the statistical analysis. They are also grateful to Calea Limited, Marchese Pharmacy (Hamilton, Ontario) and the Hamilton, Halton, and Niagara Community Care Access Centres for facilitating access to subject charts and records. Received for publication October 29, 2004. Accepted for publication April 4, 2005.
Journal of Parenteral and Enteral Nutrition, Vol. 29, No. 4,
266-269 (2005) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Top
METHODS
RESULTS
55 years vs
<55 years) and indication for PN (malignant vs other) using
unpaired t tests. 
