Cost-effectiveness Analysis of the Elder-Friendly Approaches to the Surgical Environment (EASE) Intervention for Emergency Abdominal Surgical Care of Adults Aged 65 Years and Older | Emergency Medicine | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Participant Flow Diagram
Participant Flow Diagram

The 675 participants included in the analyses were identified through index admission, and their resource use data were captured. Of those patients, 331 completed the Resource Use Inventory and EuroQol Five-Dimensions Three-Levels instrument at 6 months of follow-up. EASE indicates Elder-Friendly Approaches to the Surgical Environment.

Figure 2.  Utility, as Measured With the EuroQol Five-Dimensions Three-Levels (EQ-5D-3L) Instrument and Converted With the Canadian Scoring Algorithm
Utility, as Measured With the EuroQol Five-Dimensions Three-Levels (EQ-5D-3L) Instrument and Converted With the Canadian Scoring Algorithm

EASE indicates Elder-Friendly Approaches to the Surgical Environment.

Table 1.  Participant Characteristics at Enrollment
Participant Characteristics at Enrollment
Table 2.  Observed Resource Use and Costsa
Observed Resource Use and Costsa
Table 3.  Resource Use Inventory Patient Expenses
Resource Use Inventory Patient Expenses
1.
Government of Canada. Action for seniors report. Published October 3, 2016. Accessed November 15, 2018. https://www.canada.ca/en/employment-social-development/programs/seniors-action-report.html#tc2a
2.
Turrentine  FE, Wang  H, Simpson  VB, Jones  RS.  Surgical risk factors, morbidity, and mortality in elderly patients.  J Am Coll Surg. 2006;203(6):865-877. doi:10.1016/j.jamcollsurg.2006.08.026PubMedGoogle ScholarCrossref
3.
Gelsomino  S, Lorusso  R, Livi  U,  et al.  Cost and cost-effectiveness of cardiac surgery in elderly patients.  J Thorac Cardiovasc Surg. 2011;142(5):1062-1073. doi:10.1016/j.jtcvs.2011.02.013PubMedGoogle ScholarCrossref
4.
Rockwood  K, Howlett  SE, MacKnight  C,  et al.  Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian study of health and aging.  J Gerontol A Biol Sci Med Sci. 2004;59(12):1310-1317. doi:10.1093/gerona/59.12.1310PubMedGoogle ScholarCrossref
5.
Etzioni  DA, Liu  JH, Maggard  MA, Ko  CY.  The aging population and its impact on the surgery workforce.  Ann Surg. 2003;238(2):170-177. doi:10.1097/01.SLA.0000081085.98792.3dPubMedGoogle Scholar
6.
Creditor  MC.  Hazards of hospitalization of the elderly.  Ann Intern Med. 1993;118(3):219-223. doi:10.7326/0003-4819-118-3-199302010-00011PubMedGoogle ScholarCrossref
7.
Inouye  SK, Wagner  DR, Acampora  D, Horwitz  RI, Cooney  LM  Jr, Tinetii  ME.  A controlled trial of a nursing-centered intervention in hospitalized elderly medical patients: the Yale Geriatric Care Program.  J Am Geriatr Soc. 1993;41(12):1353-1360. doi:10.1111/j.1532-5415.1993.tb06487.xPubMedGoogle ScholarCrossref
8.
Gillick  MR, Serrell  NA, Gillick  LS.  Adverse consequences of hospitalization in the elderly.  Soc Sci Med. 1982;16(10):1033-1038. doi:10.1016/0277-9536(82)90175-7PubMedGoogle ScholarCrossref
9.
Covinsky  KE, Palmer  RM, Fortinsky  RH,  et al.  Loss of independence in activities of daily living in older adults hospitalized with medical illnesses: increased vulnerability with age.  J Am Geriatr Soc. 2003;51(4):451-458. doi:10.1046/j.1532-5415.2003.51152.xPubMedGoogle ScholarCrossref
10.
Ahmed  NN, Pearce  SE.  Acute care for the elderly: a literature review.  Popul Health Manag. 2010;13(4):219-225. doi:10.1089/pop.2009.0058PubMedGoogle ScholarCrossref
11.
Steele  JS.  Current evidence regarding models of acute care for hospitalized geriatric patients.  Geriatr Nurs. 2010;31(5):331-347. doi:10.1016/j.gerinurse.2010.03.003PubMedGoogle ScholarCrossref
12.
Fox  MT, Persaud  M, Maimets  I,  et al.  Effectiveness of acute geriatric unit care using acute care for elders components: a systematic review and meta-analysis.  J Am Geriatr Soc. 2012;60(12):2237-2245. doi:10.1111/jgs.12028PubMedGoogle ScholarCrossref
13.
Khadaroo  RG, Padwal  RS, Wagg  AS, Clement  F, Warkentin  LM, Holroyd-Leduc  J.  Optimizing senior’s surgical care: Elder-Friendly Approaches to the Surgical Environment (EASE) study—rationale and objectives.  BMC Health Serv Res. 2015;15(1):338. doi:10.1186/s12913-015-1001-2PubMedGoogle ScholarCrossref
14.
Rockwood  K, Song  X, MacKnight  C,  et al.  A global clinical measure of fitness and frailty in elderly people.  CMAJ. 2005;173(5):489-495. doi:10.1503/cmaj.050051PubMedGoogle ScholarCrossref
15.
Alberta Health. Drug products and pricing for Alberta Drug Benefit List (ADBL), Alberta Human Services Drug Benefit Supplement (HSDBS) and Palliative Coverage Drug Benefit Supplement (PCDBS). Published July 2018. Accessed November 28, 2018. https://www.ab.bluecross.ca/dbl/publications.html
16.
Statistics Canada. The consumer price index: January 2018. Published February 2018. Accessed July 9, 2018. https://www150.statcan.gc.ca/n1/pub/62-001-x/62-001-x2018001-eng.pdf
17.
EuroQol Research Foundation. EQ-5D-3L user guide. Published April 2015. Accessed June 10, 2017. https://euroqol.org/publications/user-guides
18.
Bansback  N, Tsuchiya  A, Brazier  J, Anis  A.  Canadian valuation of EQ-5D health states: preliminary value set and considerations for future valuation studies.  PLoS One. 2012;7(2):e31115. doi:10.1371/journal.pone.0031115PubMedGoogle Scholar
19.
Shoop  SJW.  Should we ban the use of “last observation carried forward” analysis in epidemiological studies?  SM J Public Health Epidemiol. 2015;1(1):1004. Accessed March 2, 2020. https://smjournals.com/public-health-epidemiology/download.php?file=fulltext/smjphe-v1-1004.pdfGoogle Scholar
20.
Dimick  JB, Ryan  AM.  Methods for evaluating changes in health care policy: the difference-in-differences approach.  JAMA. 2014;312(22):2401-2402. doi:10.1001/jama.2014.16153PubMedGoogle ScholarCrossref
21.
Sano  M, Zhu  CW, Whitehouse  PJ,  et al.  ADCS Prevention Instrument Project: pharmacoeconomics–assessing health-related resource use among healthy elderly.  Alzheimer Dis Assoc Disord. 2006;20(4)(suppl 3):S191-S202. doi:10.1097/01.wad.0000213875.63171.87PubMedGoogle ScholarCrossref
22.
Alberta Health. Alberta aids to daily living: program manual and product lists. Accessed August 23, 2018. https://www.alberta.ca/aadl-program-manual-and-product-lists.aspx
23.
Alberta Government. 2017 Alberta wage and salary survey. Published 2017. Accessed August 23, 2018. https://work.alberta.ca/documents/wage-and-salary-survey-overview.pdf
24.
Government of Canada. Consumer price index, annual average, not seasonally adjusted. Published May 30, 2018. Accessed July 9, 2018. https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1810000501
25.
Fletcher  D, MacKenzie  D, Villouta  E.  Modelling skewed data with many zeros: a simple approach combining ordinary and logistic regression.  Environ Ecol Stat. 2005;12(1):45-54. doi:10.1007/s10651-005-6817-1Google ScholarCrossref
26.
Alberta Health Services. About AHS. Accessed November 15, 2018. https://www.albertahealthservices.ca/about/about.aspx
27.
Imai  K, Kim  IS. On the use of two-way fixed effects regression models for causal inference with panel data. Published January 12, 2020. Accessed March 2, 2020. http://web.mit.edu/insong/www/pdf/FEmatch-twoway.pdf
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Health Policy
    April 3, 2020

    Cost-effectiveness Analysis of the Elder-Friendly Approaches to the Surgical Environment (EASE) Intervention for Emergency Abdominal Surgical Care of Adults Aged 65 Years and Older

    Author Affiliations
    • 1Department of Community Health Sciences, University of Calgary, Calgary, Alberta, Canada
    • 2O’Brien Institute for Public Health, University of Calgary, Calgary, Alberta, Canada
    • 3Department of Surgery, University of Alberta, Edmonton, Alberta, Canada
    • 4Department of Medicine, University of Calgary, Calgary, Alberta, Canada
    • 5Department of Medicine, University of Alberta, Edmonton, Alberta, Canada
    JAMA Netw Open. 2020;3(4):e202034. doi:10.1001/jamanetworkopen.2020.2034
    Key Points español 中文 (chinese)

    Question  From the perspective of the health care system, what are the economic outcomes of the Elder-Friendly Approaches to the Surgical Environment intervention?

    Findings  In this economic evaluation of a controlled before-and-after study of elderly Canadian adults admitted for emergency abdominal surgery, the Elder-Friendly Approaches to the Surgical Environment intervention was associated with a reduction in costs and no change in quality-adjusted life-years.

    Meaning  These findings suggest that implementation of the Elder-Friendly Approaches to the Surgical Environment intervention should be considered in other centers; in locations lacking current capacity to implement this intervention, costs to increase capacity should be weighed against the estimated costs avoided.

    Abstract

    Importance  The Elder-Friendly Approaches to the Surgical Environment (EASE) initiative is a novel approach to acute surgical care for elderly patients.

    Objective  To determine the cost-effectiveness of EASE.

    Design, Setting, and Participants  An economic evaluation from the perspective of the health care system was conducted as part of the controlled before-and-after EASE study at 2 tertiary care centers, the University of Alberta Hospital and Foothills Medical Centre. Participants included elderly adults (aged ≥65 years) admitted for emergency abdominal surgery between 2014 and 2017. Data were analyzed from April 2018 to February 2019.

    Main Outcomes and Measures  Data were captured at both control and intervention sites before and after implementation of the EASE intervention. Resource use was captured over 6 months of follow-up and was converted to costs. Utility was measured with the EuroQol Five-Dimensions Three-Levels instrument at 6 weeks and 6 months of follow-up. The differences-in-differences method was used to estimate the association of the intervention with cost and quality-adjusted life-years. For a subset of participants, self-reported out-of-pocket health care costs were collected using the Resource Use Inventory at 6 months.

    Results  A total of 675 participants were included (mean [SD] age, 75.3 [7.9] years; 333 women [49.3%]), 289 in the intervention group and 386 in the control group. The mean (SD) cost per control participant was $36 995 ($44 169) before EASE and $35 032 ($43 611) after EASE (all costs are shown in 2018 Canadian dollars). The mean (SD) cost per intervention participant was $56 143 ($74 039) before EASE and $39 001 ($59 854) after EASE. Controlling for age, sex, and Clinical Frailty Score, the EASE intervention was associated with a mean (SE) cost reduction of 23.5% (12.5%) (P = .02). The change in quality-adjusted life-years observed associated with the intervention was not statistically significant (mean [SE], 0.00001 [0.0001] quality-adjusted life-year; P = .72). The Resource Use Inventory was collected for 331 participants. The mean (SE) odds ratio for having 0 out-of-pocket expenses because of the intervention, compared with having expenses greater than 0, was 15.77 (3.37) (P = .02). Among participants with Resource Use Inventory costs greater than 0, EASE was not associated with a change in spending (mean [SE] reduction associated with EASE, 19.1% [45.2%]; P = .57).

    Conclusions and Relevance  This study suggests that the EASE intervention was associated with a reduction in costs and no change in quality-adjusted life-years. In locations that lack capacity to implement this intervention, costs to increase capacity should be weighed against the estimated costs avoided.

    Introduction

    In 2014, 15.6% of the Canadian population was aged 65 years or older, and this proportion is expected to increase to 23% by 2030.1 Age has been identified as an independent variable associated with surgical mortality and morbidity and has been associated with increased cost.2,3 This is primarily associated with the increased prevalence of frailty with age.4

    If age-specific use of surgical services remains constant, an aging population will result in a substantially increased demand for surgical services and health care costs.5 The hospital environment also subjects the older adult to increased risk of depression, delirium, and functional decline,6-8 which may not improve after hospital discharge.7,9 Therefore, measures to address hospitalization-related risks and surgical complications among older adults, as well as their associated costs, are critical.

    The Acute Care for the Elderly (ACE) model of care is an initiative developed to address some of the risks associated with hospitalization for older patients. The ACE model promotes practices of specialized patient environments, patient-centered care, medical review, and interdisciplinary team plans of care in order to prevent the commonly observed functional decline among older hospitalized medical patients.10 The use of ACE models in medical wards has been associated with decreased cost and length of hospital stay, fewer readmissions, and improved cognition, function, and satisfaction for both patients and staff.10-12 The Elder-Friendly Approaches to the Surgical Environment (EASE) initiative (ClinicalTrials.gov identifier NCT02233153) represents a novel implementation of the ACE model for acute surgical patients. The EASE initiative is a multicomponent intervention designed to enhance perioperative elderly care through capacity realignment, multidisciplinary care teams that include geriatricians, evidence-informed elder-friendly practices, functional reconditioning, and transition optimization.13

    This analysis describes findings from an economic evaluation of the EASE initiative. An economic evaluation was incorporated a priori into the overall study to assess the value of the intervention.13 Contributing to the complete reporting of all preestablished EASE study outcomes, the objective of this study is to complete an economic evaluation of EASE interventions for elderly patients recovering from emergency abdominal surgery, from the perspective of the health care system.

    Methods
    Enrollment and Follow-up

    The EASE study was a concurrently controlled before-and-after study at 2 tertiary care hospitals in Canada, which enrolled participants from 2014 to 2017. The EASE study was implemented at the University of Alberta Hospital (Edmonton, Alberta), and a comparable site, the Foothills Medical Centre (Calgary, Alberta), served as a control against temporal changes. Both hospitals are tertiary care centers and serve similar patient populations. After a period of preintervention data collection at both sites (before EASE), the EASE initiatives were implemented at the intervention site, and a comparable site with approximately the same patient intake volume and acuity level had no implementation and served as the control site (after EASE). Older adults, aged 65 years or older, admitted for emergency abdominal surgery at either study site were considered for enrollment. Only index admissions were eligible for inclusion. Readmissions, elective or trauma surgery cases, palliative interventions, and transfers from another jurisdiction or other hospital services were excluded. Patients with preoperative dependency in 3 or more activities of daily living were ineligible for study inclusion. Demographic information, including the revised Canadian Study of Health and Aging Clinical Frailty Score,14 and in-hospital preoperative, perioperative, and postoperative details were collected through medical record review.

    Before study enrollment, all participants provided written informed consent. Study participants also provided written informed consent for follow-up data collection after discharge. All EASE study protocols have been reviewed and approved at both study sites (University of Alberta Health Research Ethics Board and University of Calgary Conjoint Research Ethics). This study follows the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline.

    Economic Evaluation

    The perspective of a publicly funded health care system was taken. The primary outcome is the incremental cost-effectiveness ratio. Costs are calculated over 6 months of follow-up. Quality-adjusted life-years (QALYs) were calculated using the area-under-the-curve method with EuroQol Five-Dimensions Three-Levels (EQ-5D-3L) measurements at 4 to 8 weeks (collectively referred to as 6 weeks) and 6 months after discharge. In a secondary analysis, a broader perspective was adopted, and patient-borne costs were included.

    Resource Use and Costs

    Resource use was captured through individual, patient-level data in administrative databases for 6 months after enrollment. Index admissions and subsequent hospital admissions were captured in the Discharge Abstract Database, and in-hospital microcosting data were available from each site. For admissions with both microcosting data and Discharge Abstract Database data, microcosting was used preferentially. To calculate the costs of admissions captured in the Discharge Abstract Database only, resource intensity weight was multiplied by the cost of a standard hospital stay to arrive at the cost per episode of care. Ambulatory care was captured in the National Ambulatory Care Reporting System. Again, resource intensity weight was multiplied by the cost of a standard hospital stay to arrive at the cost per episode of care. Where the cost of a standard hospital stay was required for Discharge Abstract Database and National Ambulatory Care Reporting System observations, ordinary least-squares regression was fit to all available previous cost of a standard hospital stay to estimate values where data were not yet available. Physician billing and prescription drug costs were also included. The drug identification number for each prescription dispensed was used to identify the appropriate unit cost, which was multiplied by the number of units dispensed. The 2018 Alberta Health Drug Benefit List was used to identify the appropriate unit cost for all prescriptions dispensed.15 Where appropriate, costs have been inflated to 2018 Canadian dollars using the consumer price index.16 All costs are presented in 2018 Canadian dollars (using 2018 purchasing power parity, 1 CAD$ = $0.83 US). Because outcomes were calculated over 6 months for each participant, no discount rate was applied.

    Utility Scores

    Health-related quality of life was captured with the EQ-5D-3L instrument at 6 weeks and 6 months after discharge. The EQ-5D-3L instrument assesses self-reported health-related quality of life across 5 dimensions: mobility, self-care, usual activities, pain or discomfort, and anxiety or depression.17 Within each dimension, respondents indicate whether they have no impairment, some impairment, or complete impairment.17 The EQ-5D-3L index scores were converted to utility using the Canadian time trade-off scoring algorithm.18 Observed QALYs were calculated as area under the curve from 6 weeks to 6 months after discharge. For those participants missing 1 EQ-5D-3L response, the last observation carried forward was used to impute missing data. In this analysis, the last observation carried forward imputation assumes that in the absence of additional information, a change in utility is unlikely and avoids the loss of power associated with complete case analysis.19

    Statistical Analysis

    Data were analyzed from April 2018 to February 2019. For all analyses, statistical significance was set at 2-sided α = .05. All analyses were performed with R statistical software version 3.5.3 (R Project for Statistical Computing). The differences-in-differences (DD) method was used to analyze the association of the EASE intervention with cost and utility outcomes.20 The DD method requires a comparison group that experiences the same trends but was not exposed to the intervention.20 In the linear model used to calculate the DD, Yi = α + β1dsite + β2dintervention + β3dsite*intervention + Xi, where Yi is the outcome of interest; α is the constant term; β indicates regression coefficients; dsite is a dummy variable for site, with 1 indicating the intervention site and 0 indicating the control site; dintervention is a dummy variable for the EASE intervention, with 0 indicating before EASE and 1 indicating after EASE; dsite* intervention is the interaction term between the site and intervention variables; and Xi is a vector of controls, including age, sex, and Clinical Frailty Score. Control variables were selected on the basis of clinical relevance. Age has been identified as an independent variable associated with surgical mortality and morbidity and has been associated with increased cost.2,3 Frailty describes a combination of disease, functional ability, and other factors, such as nutritional status.4 Frailty was also selected as a control variable, to limit the impact of these features on outcomes associated with the EASE intervention. In the regression equation, the interaction term between the site and intervention variables is the DD. Cost outcomes were log-transformed before analysis with the DD method to reduce skewness and improve linearity of the association between cost and independent variables. The QALYs observed were left-skewed and required a square root transformation to reduce skewness.

    Incremental Cost-effectiveness Ratio

    Costs and QALYs at each site are calculated as those before EASE minus those after EASE. To calculate incremental costs and QALYs, pre-EASE outcomes were subtracted from post-EASE outcomes. To calculate the incremental cost-effectiveness ratio, incremental costs were divided by incremental QALYs to arrive at the additional cost to achieve 1 QALY.

    Secondary Analysis

    For a subset of study participants, self-reported costs through the Resource Utilization Inventory (RUI) were collected to capture out-of-pocket patient and caregiver costs incurred in the 6 months after discharge. The RUI has been previously validated.21 Where approved benefits through the Alberta Aids to Daily Living program existed, these costs were attached to RUI responses22 (eTable in the Supplement). The average hourly wage in Alberta in 2017 was $28.39,23 which was inflated to 2018 Canadian dollars at $29.0924 and was used to estimate costs for participant time and informal care.

    To analyze cost data collected in the RUI, a 2-stage model was used. This method was selected to accommodate data in which a substantial proportion of the values are 0 and the remainder of values have a skewed distribution.25 The RUI cost data are split into 2 data sets from the original: one indicating whether costs were 0, and the other indicating the logarithm of the cost when it was present. The presence of 0 cost observations was modeled with logistic regression, and the log-cost data were modeled with the DD method previously described.

    Results
    Participants

    A total of 684 participants were identified through an index admission; however, 9 participants were lost to follow-up before index admission cost data were collected and were excluded from this analysis. Six hundred seventy-five participants were included in this analysis: 289 in the intervention group and 386 in the control group (mean [SD] age, 75.3 [7.9] years; 333 women [49.3%]) (Figure 1). At the time of enrollment, study participants were comparable in terms of age (mean [SD] age, 75.5 [7.8] years at the control site vs 75.0 [8.0] years at the intervention site), sex (189 women [49.0%] at the control site vs 144 women [49.8%] at the intervention site), and Clinical Frailty Score (mean [SD], 3.7 [1.1] at the control site vs 4.0 [1.4] at the intervention site) (Table 1). Similar ages and proportions of each sex were observed before and after the intervention. The most common admitting diagnoses were cholecystitis or cholelithiasis (175 patients [25.9%]) and gastrointestinal obstruction not due to hernia (146 patients [21.6%]).

    At the control site, the numbers of deaths before and after EASE were similar (14 participants [8.5%] and 15 participants [6.8%], respectively). At the intervention site, 11 deaths (7.3%) were observed before EASE and 5 deaths (3.6%) were observed after EASE. Among participants who died, the mean age and Clinical Frailty Scores were similar between sites at each time point. Thirty-four participants (5.0%) lacked 1 EQ-5D-3L measurement, and responses for these participants were imputed.

    Resource Use and Costs

    The mean (SD) total cost per participant decreased over the course of 6 months from $36 995 ($44 169) to $35 032 ($43 611) at the control site, and from $56 143 ($74 039) to $39 001 ($59 854) at the intervention site (Table 2). At the control site, the mean (SD) index admission costs per participant were $23 034 ($34 762) before EASE and $24 910 ($37 285) after EASE. The mean (SD) cost of index admission at the intervention site was $33 265 ($57 722) before EASE and $25 583 ($50 893) after EASE. At the control site, the mean (SD) number of readmissions per participant were 0.3 (0.7) before EASE and 0.2 (0.5) after EASE. At the intervention site, the mean (SD) number of readmissions per participant were 0.5 (1.0) before EASE and 0.4 (0.9) after EASE. The mean (SD) length of stay per readmission decreased from 13.7 (25.5) days to 11.8 (15.1) days at the control site and from 13.8 (19.7) days to 7.5 (10.5) days at the intervention site. Among those with readmissions, the mean (SD) cost per readmission decreased from $17 667 ($31 901) to $15 350 ($15 310) at the control site and from $19 631 ($29 008) to $11 476 ($17 478) at the intervention site. The mean number of visits per participant to ambulatory care and the mean costs associated with ambulatory care were similar between sites and time points. The mean (SD) cost per participant for medications decreased from $1156 ($5247) before EASE to $888 ($3361) after EASE at the control site, and from $801 ($1396) before EASE to $751 ($1735) after EASE at the intervention site. The mean (SD) costs per participant for physician services at the control site were $4196 ($3394) before EASE and $4050 ($3991) after EASE. The mean (SD) costs per participant for physician services at the intervention site were $6893 ($7700) before the intervention and $5357 ($5778) after the intervention. Controlling for sex, age, and Clinical Frailty Score, the EASE intervention was associated with a mean (SE) cost reduction of 23.5% (12.5%) (P = .02) among all enrolled participants.

    Utility Scores

    At the control site, the mean (SD) QALY observed per participant was 0.325 (0.063) before EASE implementation and was 0.332 (0.040) after implementation of EASE. At the intervention site, the mean (SD) QALY observed per participant was 0.320 (0.058) before implementation of the intervention and was 0.329 (0.056) after implementation of the intervention (Figure 2). The change in QALYs observed associated with the intervention was not statistically significant (mean [SE], 0.00001 [0.0001] QALY; P = .72).

    Incremental Cost-effectiveness Ratio

    The mean cost per participant at the intervention site was $15 179 lower than at the control site (calculated as after EASE minus before EASE, at each site). Because the intervention site had greater QALYs observed at a reduced cost, the EASE intervention dominated the control. Density plots of QALYs and costs are included in the eFigure in the Supplement.

    Secondary Analysis

    Costs to participants are included in Table 3. Of the 331 participants for whom RUI inventory data were collected, 29 participants had costs of 0. Of the participants who had no costs in the RUI, 22 were at the control site before EASE, 3 were at the control site after EASE, 2 were at the intervention site before EASE, and 2 were at the intervention site after EASE. Before EASE, the mean (SD) cost per participant was $5091 ($9871) at the control site and $6013 ($6812) at the intervention site. After EASE, the mean (SD) cost per participant was $5458 ($9256) at the control site and $8115 ($15 808) at the intervention site. The mean (SE) odds ratio of having 0 expenses captured by the RUI because of the EASE intervention, compared with having expenses greater than 0, was 15.77 (3.37) (P = .02) controlling for age, sex, and Clinical Frailty Score. Among those participants with RUI costs greater than 0, the EASE intervention was not associated with a statistically significant change in spending (mean [SE] reduction associated with EASE, 19.1% [45.2%]; P = .57).

    Discussion

    From the health care system perspective, the EASE initiative was associated with a substantial reduction in costs, with no evidence of a difference in QALY outcomes. Costs attached to self-reported RUI responses suggested that patients and their caregivers were less likely to spend money on health care when EASE interventions were implemented. These findings indicate that elder-friendly perioperative care may reduce costs to the health care system, without increasing costs to patients and without detriment to effectiveness measured in QALYs.

    The context of the EASE study13 provided an ideal laboratory for economic evaluation. The DD method assumes that in the absence of the intervention, the unobserved differences between control and intervention groups are the same over time. The tertiary care centers selected as intervention and control sites in this study were selected because they had similar patient populations served in terms of volume, acuity, and proximity to university medical schools. Alberta stands out among Canadian provinces for having the largest provincewide, fully integrated health system.26 Provincial coordination of health care services reduces heterogeneity in care provided and supports the main assumption of DD. Furthermore, provincial collection of administrative data used to inform costing eliminates loss of data that may occur because of differences in data collection across other health systems.

    Interestingly, mean costs per participant at the intervention site were greater than those at the control site at each time point, which suggests the presence of unmeasured systematic differences between the 2 sites. In this study, data were collected from 2 groups at 2 time points, with 1 group at the last time point receiving the intervention. This is the simplest setting for use of the DD method and, in this case, is equivalent to the 2-way fixed-effects approach with treatment and time indicators.27 If a greater number of study sites had been included, the 2-way fixed-effects approach could have been used to reduce the influence of preperiod-level differences on estimates of effect. Additional work is required to understand differences between the 2 sites and to take advantage of cost-reducing practices occurring already at the control site.

    In a systematic review by Ahmed and Pearce,10 6 studies reported costs of ACE models of care. In 2 of these studies, a neutral outcome on cost was reported.10 In the remaining 4 studies, the ACE model was associated with a decrease in total costs, hospital costs, or cost of care.10 This evaluation of an ACE model in the surgical setting supports the generalization of cost reductions identified in previous ACE studies to the setting of emergency abdominal surgical patients.

    Limitations

    This study has several limitations. First, this was not a randomized clinical trial, which limits the ability to infer causation. Second, the RUIs are affected by a substantial loss to follow-up, which reduces the power to detect effects of EASE interventions. Compared with RUI responders, those who were lost to follow-up had much greater mean clinical frailty scores and greater mean age, which likely contributed to patient health care spending. However, the comparable numbers of participants lost to follow-up at each site and at each time point reduces the potential impact of this limitation. Third, implementation costs associated with the EASE interventions were not included in this analysis; therefore, initial costs of care cannot be commented on. In settings that already have the network of multidisciplinary care practitioners required to implement EASE, the increased coordination of resources would not significantly change workflow. This case involved 2 well-resourced tertiary care settings in large cities capable of implementing EASE without additional cost. Of note, no new staff were hired to implement EASE. Instead, grant funding was used to hire research assistants for data collection.

    Conclusions

    In the setting of acute abdominal surgery, elder-friendly approaches to care have been associated with reduced health care system costs and were not associated with QALYs. Patients and their caregivers were also less likely to spend money on health services after EASE interventions. This intervention should be implemented broadly. In locations that do not already have the capacity to implement this intervention, costs to implement should be weighed against the estimated costs avoided.

    Back to top
    Article Information

    Accepted for Publication: February 4, 2020.

    Published: April 3, 2020. doi:10.1001/jamanetworkopen.2020.2034

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Hofmeister M et al. JAMA Network Open.

    Corresponding Author: Fiona Clement, PhD, Department of Community Health Sciences, University of Calgary, 3280 Hospital Dr NW, 3D14 Teaching Research and Wellness Bldg, Calgary, AB T2N 4N1, Canada (fclement@ucalgary.ca).

    Author Contributions: Mr Hofmeister and Dr Clement had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Khadaroo, Holroyd-Leduc, Padwal, Wagg, Warkentin, Clement.

    Acquisition, analysis, or interpretation of data: Hofmeister, Holroyd-Leduc, Padwal, Warkentin, Clement.

    Drafting of the manuscript: Hofmeister, Wagg, Clement.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Hofmeister.

    Obtained funding: Khadaroo, Holroyd-Leduc, Clement.

    Administrative, technical, or material support: Padwal, Wagg, Warkentin.

    Supervision: Holroyd-Leduc, Padwal, Wagg, Clement.

    Conflict of Interest Disclosures: Dr Holroyd-Leduc reported receiving grants from Alberta Innovates–Health Solutions and Alberta Health Services during the conduct of the study. Dr Wagg reported receiving grants and personal fees from Astellas Pharma and Essity Health and Hygiene AB; personal fees from Pierre Fabre Medicamens; and grants, personal fees, and nonfinancial support from Pfizer outside the submitted work. No other disclosures were reported.

    Funding/Support: This work was funded by a Partnership for Research and Innovation in the Health System grant from Alberta Innovates.

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Additional Contributions: Ashley Wanamaker, MC (University of Calgary), Carrie Le, MC (University of Calgary), Hanhmi Huynh, BSN, BSc (University of Alberta), Jenelle Pederson, MSc (University of Calgary), and Justin Lowes, BSc (University of Alberta), assisted with data collection. All of them were financially compensated through the course of their employment for their contributions. Alberta Innovates, the Surgery & Seniors Health Strategic Clinical Networks, and the Acute Care Surgery Services, Alberta Health Services provided nonfinacial support for the EASE project in terms of managing human resources and championing the intervention.

    References
    1.
    Government of Canada. Action for seniors report. Published October 3, 2016. Accessed November 15, 2018. https://www.canada.ca/en/employment-social-development/programs/seniors-action-report.html#tc2a
    2.
    Turrentine  FE, Wang  H, Simpson  VB, Jones  RS.  Surgical risk factors, morbidity, and mortality in elderly patients.  J Am Coll Surg. 2006;203(6):865-877. doi:10.1016/j.jamcollsurg.2006.08.026PubMedGoogle ScholarCrossref
    3.
    Gelsomino  S, Lorusso  R, Livi  U,  et al.  Cost and cost-effectiveness of cardiac surgery in elderly patients.  J Thorac Cardiovasc Surg. 2011;142(5):1062-1073. doi:10.1016/j.jtcvs.2011.02.013PubMedGoogle ScholarCrossref
    4.
    Rockwood  K, Howlett  SE, MacKnight  C,  et al.  Prevalence, attributes, and outcomes of fitness and frailty in community-dwelling older adults: report from the Canadian study of health and aging.  J Gerontol A Biol Sci Med Sci. 2004;59(12):1310-1317. doi:10.1093/gerona/59.12.1310PubMedGoogle ScholarCrossref
    5.
    Etzioni  DA, Liu  JH, Maggard  MA, Ko  CY.  The aging population and its impact on the surgery workforce.  Ann Surg. 2003;238(2):170-177. doi:10.1097/01.SLA.0000081085.98792.3dPubMedGoogle Scholar
    6.
    Creditor  MC.  Hazards of hospitalization of the elderly.  Ann Intern Med. 1993;118(3):219-223. doi:10.7326/0003-4819-118-3-199302010-00011PubMedGoogle ScholarCrossref
    7.
    Inouye  SK, Wagner  DR, Acampora  D, Horwitz  RI, Cooney  LM  Jr, Tinetii  ME.  A controlled trial of a nursing-centered intervention in hospitalized elderly medical patients: the Yale Geriatric Care Program.  J Am Geriatr Soc. 1993;41(12):1353-1360. doi:10.1111/j.1532-5415.1993.tb06487.xPubMedGoogle ScholarCrossref
    8.
    Gillick  MR, Serrell  NA, Gillick  LS.  Adverse consequences of hospitalization in the elderly.  Soc Sci Med. 1982;16(10):1033-1038. doi:10.1016/0277-9536(82)90175-7PubMedGoogle ScholarCrossref
    9.
    Covinsky  KE, Palmer  RM, Fortinsky  RH,  et al.  Loss of independence in activities of daily living in older adults hospitalized with medical illnesses: increased vulnerability with age.  J Am Geriatr Soc. 2003;51(4):451-458. doi:10.1046/j.1532-5415.2003.51152.xPubMedGoogle ScholarCrossref
    10.
    Ahmed  NN, Pearce  SE.  Acute care for the elderly: a literature review.  Popul Health Manag. 2010;13(4):219-225. doi:10.1089/pop.2009.0058PubMedGoogle ScholarCrossref
    11.
    Steele  JS.  Current evidence regarding models of acute care for hospitalized geriatric patients.  Geriatr Nurs. 2010;31(5):331-347. doi:10.1016/j.gerinurse.2010.03.003PubMedGoogle ScholarCrossref
    12.
    Fox  MT, Persaud  M, Maimets  I,  et al.  Effectiveness of acute geriatric unit care using acute care for elders components: a systematic review and meta-analysis.  J Am Geriatr Soc. 2012;60(12):2237-2245. doi:10.1111/jgs.12028PubMedGoogle ScholarCrossref
    13.
    Khadaroo  RG, Padwal  RS, Wagg  AS, Clement  F, Warkentin  LM, Holroyd-Leduc  J.  Optimizing senior’s surgical care: Elder-Friendly Approaches to the Surgical Environment (EASE) study—rationale and objectives.  BMC Health Serv Res. 2015;15(1):338. doi:10.1186/s12913-015-1001-2PubMedGoogle ScholarCrossref
    14.
    Rockwood  K, Song  X, MacKnight  C,  et al.  A global clinical measure of fitness and frailty in elderly people.  CMAJ. 2005;173(5):489-495. doi:10.1503/cmaj.050051PubMedGoogle ScholarCrossref
    15.
    Alberta Health. Drug products and pricing for Alberta Drug Benefit List (ADBL), Alberta Human Services Drug Benefit Supplement (HSDBS) and Palliative Coverage Drug Benefit Supplement (PCDBS). Published July 2018. Accessed November 28, 2018. https://www.ab.bluecross.ca/dbl/publications.html
    16.
    Statistics Canada. The consumer price index: January 2018. Published February 2018. Accessed July 9, 2018. https://www150.statcan.gc.ca/n1/pub/62-001-x/62-001-x2018001-eng.pdf
    17.
    EuroQol Research Foundation. EQ-5D-3L user guide. Published April 2015. Accessed June 10, 2017. https://euroqol.org/publications/user-guides
    18.
    Bansback  N, Tsuchiya  A, Brazier  J, Anis  A.  Canadian valuation of EQ-5D health states: preliminary value set and considerations for future valuation studies.  PLoS One. 2012;7(2):e31115. doi:10.1371/journal.pone.0031115PubMedGoogle Scholar
    19.
    Shoop  SJW.  Should we ban the use of “last observation carried forward” analysis in epidemiological studies?  SM J Public Health Epidemiol. 2015;1(1):1004. Accessed March 2, 2020. https://smjournals.com/public-health-epidemiology/download.php?file=fulltext/smjphe-v1-1004.pdfGoogle Scholar
    20.
    Dimick  JB, Ryan  AM.  Methods for evaluating changes in health care policy: the difference-in-differences approach.  JAMA. 2014;312(22):2401-2402. doi:10.1001/jama.2014.16153PubMedGoogle ScholarCrossref
    21.
    Sano  M, Zhu  CW, Whitehouse  PJ,  et al.  ADCS Prevention Instrument Project: pharmacoeconomics–assessing health-related resource use among healthy elderly.  Alzheimer Dis Assoc Disord. 2006;20(4)(suppl 3):S191-S202. doi:10.1097/01.wad.0000213875.63171.87PubMedGoogle ScholarCrossref
    22.
    Alberta Health. Alberta aids to daily living: program manual and product lists. Accessed August 23, 2018. https://www.alberta.ca/aadl-program-manual-and-product-lists.aspx
    23.
    Alberta Government. 2017 Alberta wage and salary survey. Published 2017. Accessed August 23, 2018. https://work.alberta.ca/documents/wage-and-salary-survey-overview.pdf
    24.
    Government of Canada. Consumer price index, annual average, not seasonally adjusted. Published May 30, 2018. Accessed July 9, 2018. https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1810000501
    25.
    Fletcher  D, MacKenzie  D, Villouta  E.  Modelling skewed data with many zeros: a simple approach combining ordinary and logistic regression.  Environ Ecol Stat. 2005;12(1):45-54. doi:10.1007/s10651-005-6817-1Google ScholarCrossref
    26.
    Alberta Health Services. About AHS. Accessed November 15, 2018. https://www.albertahealthservices.ca/about/about.aspx
    27.
    Imai  K, Kim  IS. On the use of two-way fixed effects regression models for causal inference with panel data. Published January 12, 2020. Accessed March 2, 2020. http://web.mit.edu/insong/www/pdf/FEmatch-twoway.pdf
    ×