Abstract
Background
This study quantifies incremental healthcare expenditures of functional impairments and phenotypic frailty in specific healthcare sectors.
Methods
Pooled 2023 analysis of 4 prospective cohort studies linked with Medicare claims including 4 318 women and 3 847 men attending an index examination (2002–2011). Annualized inpatient, skilled nursing facility (SNF), home healthcare (HHC), and outpatient costs (2023 dollars) ascertained for 36 months following index examination. Functional impairments (difficulty performing 4 activities of daily living) and frailty phenotype (operationalized using 5 components) derived from cohort data. Weighted multimorbidity index including demographics derived from claims.
Results
Mean age at index examination was 79.2 years. After accounting for multimorbidity and each other, average annualized incremental costs of 3–4 functional impairments versus no impairment in women (men) were $2 838 ($5 516) in inpatient, $1 572 ($1 446) in SNF, and $1 349 ($1 060) in HHC sectors; average incremental costs of phenotypic frailty versus robust in women (men) was $4 100 (not significant for men) in inpatient, $1 579 ($1 254) in SNF, and $645 ($526) in HHC sectors. Incremental inpatient costs were primarily due to a higher hospitalization risk, while incremental SNF and HHC costs were related to both increased risks of utilization and higher costs among individuals with utilization. Neither geriatric domain was associated with outpatient costs.
Conclusions
In this study of community-dwelling beneficiaries, functional impairments were independently associated with higher subsequent expenditures in inpatient, SNF, and HHC sectors among both sexes. Phenotypic frailty was independently associated with higher subsequent inpatient costs in women, and higher SNF and HHC costs in both sexes.
Keywords: Frail, Functional status, Health services, Medicare claims data
Older adults account for a disproportionate share of healthcare expenditures and utilization. Average total healthcare expenditure is 3 times higher for adults aged 65 years and older compared to those aged 19–64 years (1). While a minority of older adults account for most of the healthcare costs in this age group (2), they comprise a diverse group of individuals (3). Conventional models for prediction of healthcare costs primarily rely on claims-based indicators of multimorbidity measuring the number and complexity of medical conditions (4,5). However, these models account for only a modest proportion of the variation in individual expenditures in community-dwelling older Medicare beneficiaries because of failure to adequately account for differences in patient characteristics (6). Thus, better strategies are needed by healthcare systems to identify and depict vulnerable older adults at increased risk for costly care to enhance selection of target groups for interventions aimed at reducing future spending (7).
A growing body of literature indicates that the interrelated but distinct geriatric domains of self-reported functional impairments and phenotypic frailty (8,9) are risk factors for higher total healthcare costs and utilization independent of multimorbidity (10–17). Neither domain is routinely assessed in time-limited or resource-constrained practice settings such as primary care clinic or captured in electronic health record or claims data, but these 2 characteristics indicate increased vulnerability to adverse health outcomes and reduced ability to recover from illness. We recently reported substantial incremental total healthcare expenditures in community-dwelling older adults attributable to functional impairments in both sexes and phenotypic prefrailty and frailty in women that were not accounted for by claims-derived cost indicators of multimorbidity and frailty or each other (10). Other studies found increased risks of acute hospitalization, skilled nursing facility (SNF) stays or readmission among older adults with functional impairments (11–15), or phenotypic frailty (16,17).
However, no previous investigation quantified incremental sector-specific costs of functional impairments or phenotypic frailty. It is also uncertain whether any incremental sector-specific expenditures among older adults with functional impairments or phenotypic frailty result from increased risks of utilization alone or are also due to higher costs among individuals receiving care in these sectors. In addition, the contribution of expenditures in specific sectors to total healthcare costs attributable to functional impairments or frailty is unknown. This information is needed by healthcare systems better inform understanding of the trade-off between resources required to assess self-reported function and the frailty phenotype in their older patients and potential cost savings in specific healthcare sectors from interventions aimed at reducing future spending by slowing progression to overt disability and severe frailty.
To determine whether functional impairments and phenotypic frailty are associated with subsequent incremental healthcare costs in inpatient, SNF, home healthcare (HHC), and outpatient sectors after accounting for multimorbidity and each other, we utilized a unique multi-cohort data set linked with Medicare claims data composed of 8 165 older community-dwelling adults (4 318 women and 3 847 men) who were continuously enrolled in the Medicare fee-for-service (FFS) program. In addition, we assessed whether any incremental costs of functional impairments and phenotypic frailty in these sectors were due to increased risks of utilization alone or also to higher costs among individuals with utilization. We also determined the contribution of sector-specific expenditures to total healthcare costs attributable to functional impairments and frailty.
Method
Participants
We studied community-dwelling older adults enrolled in 4 prospective cohort studies who completed a comprehensive in-person index examination including assessment of self-reported functional impairments and the frailty phenotype: Year 16 (2002–2004) examination of the Study of Osteoporotic Fractures (SOF), Year 7 (2007–2009) examination of the Osteoporotic Fracture in Men Study (MrOS), Year 6 (2002–2003) examination of the Health, Aging and Body Composition Study (Health ABC), and the 2011 examination of the National Health and Aging Trends Study (NHATS; Supplementary Table 1). Of the 8 393 women and 7 653 men with complete cohort study data at the index examination (Supplementary Figure 1), our analytical cohorts included 4 318 women and 3 847 men who were also enrolled in the Medicare FFS program from 12 months prior to the index examination until 36 months following the index examination (or until death within this follow-up period). FFS participants who died within 36 months following the index examination were included in analyses; their costs were annualized based on the number of days they were alive during the 36-month follow-up period. The University of Minnesota IRB approved the study; informed consent was waived because it was a secondary analysis of deidentified data. We followed Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines.
Healthcare Costs and Utilization
Standardized annualized total direct and sector-specific healthcare costs for 36 months following the index examination were calculated from the healthcare sector perspective (18,19) and included costs paid by Medicare, costs paid by supplementary insurance, and out-of-pocket payments made by patients. Costs were annualized accounting for person-time (number of days alive) and adjusted to U.S. 2023 dollars (20). Total direct annualized healthcare costs were calculated as the sum of costs for acute hospital stays, inpatient rehabilitation facility (IRF) stays, SNF stays paid at least in part under Medicare part A, outpatient care, durable medical equipment, and HHC for that time period (20–22). Primary outcomes of this analysis were annualized costs in inpatient, SNF, HHC, and outpatient sectors for 36 months following the index examination. Inpatient costs for each participant during the follow-up period were calculated as the sum of costs due to hospitalizations or IRF stays, identified from either the MedPAR or Inpatient SAF files, plus the inpatient carrier claims, identified by place of service within the Carrier files. SNF costs were identified from either MedPAR or SNF SAF files. HHC costs were from the HHC files. Outpatient costs were the sum of costs from the outpatient files and the outpatient carrier claims, identified by place of service within the Carrier files.
Functional Impairments and Frailty Phenotype
Measures of functional impairments and the frailty phenotype at the index examination were harmonized as previously described (Supplementary Tables 2 and 3) (10). Participants were asked about difficulty performing 4 activities of daily living (ADL; ie, walking a few blocks on level ground, climbing up 10–20 steps, transferring from bed to chair, and bathing or showering). Participants were categorized by number of functional impairments. The frailty phenotype was assessed for each participant using a modified version of 5 components initially proposed by Fried and colleagues (23) including shrinkage (recent weight loss of ≥5% or ≥10 pounds or body mass index <18.5 kg/m2) (17), weakness (grip strength <32 kg for men or <20 kg for women) (24), self-reported poor energy, slowness (gait speed <0.8 m/s for men or <0.6 m/s for women or use of a walking aid) (17,25,26), and low physical activity (never walking for exercise and not engaging in moderate or vigorous activity) (27). Participants were classified as robust (no components), prefrail (1 or 2 components), or frail (≥3 components).
Multimorbidity
A weighted claims-based multimorbidity measure (community version 12 [v1209.F1] of the Centers for Medicare & Medicaid Services Hierarchical Conditions Categories [CMS-HCC] model) (28) was calculated for each participant using demographic data (age, sex, Medicaid eligibility, and disability status) and over 14 000 diagnosis codes in inpatient and outpatient claims in the year prior to the index examination (5,6,29). The CMS-HCC score is specifically calibrated to predict healthcare costs in Medicare beneficiaries.
Other Measurements
The Kim claims-derived frailty indicator (CFI) (30) approximating the deficit accumulation index (9) was calculated using diagnosis and procedure codes in the 12 months prior to the index examination. A measure of neighborhood socioeconomic deprivation (area deprivation index [ADI] of participant residence census tract) was calculated using 2006–2010 data from the U.S. Census American Community Survey (31). Cognitive impairment was defined at the index examination by a self- or proxy report of dementia or abnormal cognitive test results.
Statistical Analysis
All analyses were performed using Stata version 17.0 and stratified by sex (StataCorp LLC, College Station, TX). Independent variables in all models included geographic region (study enrollment site for SOF, MrOS, and Health ABC and census region for NHATS) (32), CMS-HCC score (log-transformed, then standardized), number of functional impairments (none [referent group], 1, 2, 3, or 4), and the frailty phenotype (robust [referent group], prefrail, frail).
Because more than half of participants had zero costs for inpatient hospitalization (53.8%), SNF stays (86.1%), or HHC episodes (74.6%) during the 36-month follow-up period, 2-part models were employed for each of these sectors to estimate annualized costs by number of functional impairments and category of frailty phenotype. Logistic models estimated the odds of ≥1 hospitalization, SNF stay or HHC episode, and generalized linear models (33) with gamma distributions and log links estimated cost ratios (CRs) among participants with nonzero costs. The Stata twopart and margins commands were used to estimate mean annualized predicted and incremental costs for all participants (including those with zero costs) in each sector by number of functional impairments and category of frailty phenotype. Gamma distribution models with log links were used to estimate associations of functional impairments and the frailty phenotype with annualized outpatient and total healthcare costs. In a secondary analysis, we examined the contribution of sector-specific expenditures to total healthcare costs by number of functional impairments and category of frailty phenotype. We also performed sensitivity analyses further adjusting models for Kim CFI, ADI, or cognitive impairment.
Statistical significance was defined by 95% confidence interval (CI) excluding the null value when 95% CI was calculated and p < .05 when the null hypothesis test was performed.
Results
The study population of community-dwelling FFS beneficiaries included 4 318 women (mean age 80.0 years) and 3 847 men (mean age 78.4 years; Supplementary Table 4). At least 1 functional impairment was reported by 2 120 (49.1%) women and 1 227 (31.9%) men; 1 480 (34.3%) women and 733 (19.1%) men were classified as frail using the phenotypic definition. Mean (standard deviation) annualized unadjusted total healthcare expenditures during the 36-month follow-up period were $13 906 ($24 499) for women and $14 598 ($28 556) for men.
Participants reporting greater functional impairment or those with a greater degree of frailty among both sexes were older, had higher CMS-HCC scores, higher subsequent total and outpatient healthcare costs, and greater risks of hospitalization, SNF stay, HHC use, and mortality during follow-up (Tables 1–2). Among individuals with 3–4 functional impairments, 74.2% of women and 71.0% of men were frail, and over one-half of frail individuals (60.3% of women and 55.5% of men) reported ≥2 impairments.
Table 1.
Characteristics of Participants by Number of Functional Impairments
| Characteristic | Functional Impairments* | |||
|---|---|---|---|---|
| None | 1 | 2 | 3–4 | |
| Women (n = 4 318) | n = 2 198 | n = 747 | n = 694 | n = 679 |
| Age, years, mean (SD) | 78.7 (6.0) | 80.3 (6.6) | 81.6 (6.6) | 82.0 (7.3) |
| CMS-HCC score | ||||
| Median [IQR] | 0.74 [0.54, 1.10] | 1.04 [0.66, 1.58] | 1.18 [0.77, 1.86] | 1.52 [0.97, 2.44] |
| Mean (SD) | 0.90 (0.59) | 1.23 (0.76) | 1.52 (1.10) | 1.93 (1.38) |
| Phenotypic frailty, n (%) | 316 (14.4) | 271 (36.3) | 389 (56.1) | 504 (74.2) |
| Total annual healthcare costs, USD | ||||
| Median [IQR] | $3 991 [1 768, 9 776] | $7 936 [3 077, 18 471] | $10 027 [3 733, 25 787] | $16 611 [6 399, 36 383] |
| Mean (SD) | $9 019 (18 380) | $15 743 (26 424) | $19 992 (28 934) | $28 183 (37 462) |
| Hospitalization during follow-up, n (%) | 777 (35.4) | 403 (54.0) | 400 (57.6) | 476 (70.1) |
| SNF stay during follow-up, n (%) | 207 (9.4) | 147 (19.7) | 166 (23.9) | 209 (30.8) |
| HHC during follow-up, n (%) | 360 (16.4) | 241 (32.3) | 272 (39.2) | 388 (57.1) |
| Annual outpatient healthcare costs, USD | ||||
| Median [IQR] | $2 764 [1 429, 4 869] | $3 610 [2 082, 6 074] | $3 989 [2 110, 6 876] | $4 559 [2 331, 8 036] |
| Mean (SD) | $4 056 (5 105) | $5 518 (8 011) | $6 110 (8 158) | $7 468 (10 723) |
| Died during follow-up, n (%) | 128 (5.8) | 82 (11.0) | 133 (19.2) | 231 (34.0) |
| Men (n = 3 847) | n = 2 620 | n = 527 | n = 341 | n = 359 |
|---|---|---|---|---|
| Age, years, mean (SD) | 77.6 (5.6) | 79.9 (6.0) | 79.8 (6.5) | 80.9 (7.2) |
| CMS-HCC score | ||||
| Median [IQR] | 0.81 [0.52, 1.24] | 1.13 [0.69, 1.84] | 1.42 [0.86, 2.20] | 1.66 [0.99, 2.98] |
| Mean (SD) | 1.03 (0.75) | 1.45 (1.10) | 1.79 (1.29) | 2.15 (1.58) |
| Phenotypic frailty, n (%) | 183 (7.0) | 143 (27.1) | 152 (44.6) | 255 (71.0) |
| Total annual healthcare costs, USD | ||||
| Median [IQR] | $4 705 [1 975, 11 517] | $9 594 [2 765, 22 160] | $13 495 [5 138, 31 170] | $18 458 [6 424, 43 027] |
| Mean (SD) | $11 180 (24 102) | $17 536 (24 461) | $26 528 (38 025) | $35 564 (55 290) |
| Hospitalization during follow-up, n (%) | 981 (37.4) | 288 (54.7) | 214 (62.8) | 235 (65.5) |
| SNF stay during follow-up, n (%) | 153 (5.8) | 86 (16.3) | 73 (21.4) | 91 (25.4) |
| HHC during follow-up, n (%) | 339 (12.9) | 163 (30.9) | 132 (38.7) | 182 (50.7) |
| Annual outpatient healthcare costs, USD | ||||
| Median [IQR] | $3 317 [1 579, 5 951] | $4 164 [1 931, 7 946] | $4 951 [2 338, 9 097] | $5 396 [2 314, 10 093] |
| Mean (SD) | $5 213 (7 757) | $6 986 (9 692) | $8 503 (12 309) | $8 179 (9 926) |
| Died during follow-up, n (%) | 197 (7.5) | 87 (16.5) | 84 (24.6) | 138 (38.4) |
Notes: CMS-HCC = Centers for Medicare & Medicaid Services Hierarchical Condition Categories; HHC = home healthcare; IQR = interquartile range; SD = standard deviation; SNF = skilled nursing facility; USD = 2023 U.S. dollars.
*p Value <.001 for comparison of each characteristic across categories of functional impairments among women and men.
Table 2.
Characteristics of Participants by Category of Frailty Phenotype
| Category of Frailty Phenotype* | |||
|---|---|---|---|
| Characteristic | Robust | Prefrail | Frail |
| Women (n = 4 318) | n = 724 | n = 2 114 | n = 1 480 |
| Age, years, mean (SD) | 77.0 (5.5) | 79.3 (6.3) | 82.4 (6.6) |
| CMS-HCC score | |||
| Median [IQR] | 0.58 [0.46, 0.97] | 0.84 [0.54, 1.28] | 1.29 [0.85, 2.01] |
| Mean (SD) | 0.80 (0.55) | 1.05 (0.69) | 1.67 (1.23) |
| At least 2 functional impairments, n (%) | 49 (6.8) | 431 (20.4) | 893 (60.3) |
| Total annual healthcare costs, USD | |||
| Median [IQR] | $2 966 [1 420, 6 830] | $5 376 [2 236, 13 051] | $11 845 [4 647, 29 761] |
| Mean (SD) | $7 506 (16 863) | $11 502 (21 159) | $23 544 (33 732) |
| Hospitalization during follow-up, n (%) | 186 (25.7) | 910 (43.1) | 960 (64.9) |
| SNF stay during follow-up, n (%) | 33 (4.6) | 267 (12.6) | 429 (29.0) |
| HHC during follow-up, n (%) | 85 (11.7) | 483 (22.8) | 693 (46.8) |
| Annual outpatient healthcare costs, USD | |||
| Median [IQR] | $2 402 [1 271, 4 271] | $3 208 [1 718, 5 589] | $4 043 [2 127, 7 213] |
| Mean (SD) | $3 548 (4 663) | $4 791 (6 513) | $6 522 (9 229) |
| Died during follow-up, n (%) | 31 (4.3) | 164 (7.8) | 379 (25.6) |
| Men (n = 3 847) | n = 1 109 | n = 2 005 | n = 733 |
|---|---|---|---|
| Age, years, mean (SD) | 76.9 (5.2) | 78.3 (5.8) | 80.8 (6.9) |
| CMS-HCC score | |||
| Median [IQR] | 0.73 [0.52, 1.08] | 0.90 [0.60, 1.50] | 1.57 [0.91, 2.65] |
| Mean (SD) | 0.90 (0.60) | 1.19 (0.89) | 2.00 (1.47) |
| At least 2 functional impairments, n (%) | 23 (2.1) | 270 (13.5) | 407 (55.5) |
| Total annual healthcare costs, USD | |||
| Median [IQR] | $4 300 [1 776, 9 888] | $5 769 [2 303, 15 263] | $14 825 [5 153, 36 498] |
| Mean (SD) | $9 824 (19 553) | $13 704 (27 720) | $29 983 (45 264) |
| Hospitalization during follow-up, n (%) | 408 (36.8) | 863 (43.0) | 447 (61.0) |
| SNF stay during follow-up, n (%) | 46 (4.2) | 193 (9.6) | 164 (22.4) |
| HHC during follow-up, n (%) | 113 (10.2) | 389 (19.4) | 314 (42.8) |
| Annual outpatient healthcare costs, USD | |||
| Median [IQR] | $3 023 [1 523, 5 410] | $3 599 [1 651, 6 584] | $5 377 [2 476, 9 572] |
| Mean (SD) | $5 048 (9 003) | $5 534 (7 116) | $8 843 (11 751) |
| Died during follow-up, n (%) | 63 (5.7) | 215 (10.7) | 228 (31.1) |
Notes: CMS-HCC = Centers for Medicare & Medicaid Services Hierarchical Condition Categories; HHC = home healthcare; IQR = interquartile range; SD = standard deviation; SNF = skilled nursing facility; USD = 2023 U.S. dollars.
*p Value <.001 for comparison of each characteristic across categories of frailty phenotype among women and men.
Functional Impairments, Phenotypic Frailty, and Inpatient Costs
In a model including geographic region, CMS-HCC score, functional impairments, and the frailty phenotype as independent variables (multivariable model), mean predicted inpatient expenditures among individuals without impairments were $5 235 (95% CI: $4 410, $6 062) in women and $5 923 (95% CI: $5 121, $6 725) in men (Table 3). Average incremental inpatient costs incurred by individuals with impairments versus no impairment ranged from $1 421 (95% CI: $37, $2 805) for 1 impairment to $2 838 (95% CI: $1 004, $4 671) for 3–4 impairments among women, and from $1 221 (not significantly different from zero, 95% CI: −$467, $2 909) for 1 impairment to $5 516 (95% CI: $2 285, $8 747) for 3–4 impairments among men. Additional costs among those with versus without impairments were largely driven by higher odds of hospitalization in both sexes (odds ratio [OR] for 1 impairment 1.62 [95% CI: 1.35, 1.95] in women and 1.57 [95% CI: 1.28, 1.92] in men up to 2.62 [95% CI: 2.08, 3.30] in women and 1.86 [95% CI: 1.41, 2.45] in men for individuals with 3–4 impairments). However, among men who were hospitalized, 3–4 versus no impairments was associated with a 1.5-fold increase in inpatient costs (CR: 1.48, 95% CI: 1.10, 1.98).
Table 3.
Effects of Functional Impairments on Annualized Sector-Specific Healthcare Costs and Utilization*
| Mean Predicted Costs (95% CI) | Mean Incremental Costs (95% CI) | Odds Ratio of Utilization (95% CI) | Cost Ratio Among Those With Utilization (95% CI) | |
|---|---|---|---|---|
| Inpatient sector | ||||
| Women† | ||||
| 0 | $5 235 (4 410, 6 061) | Referent | Referent | Referent |
| 1 | $6 656 (5 456, 7 856) | $1 421 (37, 2 805) | 1.62 (1.35, 1.95) | 1.02 (0.83, 1.26) |
| 2 | $7 309 (5 986, 8 631) | $2 074 (484, 3 663) | 1.67 (1.36, 2.04) | 1.11 (0.89, 1.40) |
| 3–4 | $8 073 (6 622, 9 524) | $2 838 (1 004, 4 671) | 2.62 (2.08, 3.30) | 1.05 (0.82, 1.34) |
| Men† | ||||
| 0 | $5 923 (5 121, 6 725) | Referent | Referent | Referent |
| 1 | $7 144 (5 633, 8 655) | $1 221 (−467, 2 909) | 1.57 (1.28, 1.92) | 0.99 (0.78, 1.24) |
| 2 | $9 317 (7 001, 11 633) | $3 394 (868, 5 920) | 1.91 (1.47, 2.47) | 1.19 (0.91, 1.57) |
| 3–4 | $11 439 (8 466, 14 413) | $5 516 (2 285, 8 747) | 1.86 (1.41, 2.45) | 1.48 (1.10, 1.98) |
| SNF sector | ||||
| Women‡ | ||||
| 0 | $931 (733, 1 129) | Referent | Referent | Referent |
| 1 | $1 750 (1 374, 2 126) | $819 (412, 1 227) | 1.77 (1.38, 2.26) | 1.20 (0.97, 1.48) |
| 2 | $2 073 (1 662, 2 485) | $1 142 (670, 1 615) | 1.96 (1.51, 2.55) | 1.32 (1.05, 1.65) |
| 3–4 | $2 504 (2 022, 2 985) | $1 572 (1 010, 2 135) | 2.62 (1.97, 3.46) | 1.30 (1.03, 1.64) |
| Men‡ | ||||
| 0 | $618 (454, 781) | Referent | Referent | Referent |
| 1 | $1 257 (885, 1 629) | $639 (239, 1 040) | 2.31 (1.71, 3.12) | 1.00 (0.75, 1.34) |
| 2 | $1 529 (1 033, 2 026) | $912 (376, 1 448) | 2.68 (1.90, 3.77) | 1.08 (0.78, 1.50) |
| 3–4 | $2 064 (1 389, 2 739) | $1 446 (713, 2 179) | 2.81 (1.97, 4.01) | 1.41 (0.98, 2.01) |
| HHC sector | ||||
| Women§ | ||||
| 0 | $490 (411, 569) | Referent | Referent | Referent |
| 1 | $740 (619, 862) | $250 (114, 386) | 1.75 (1.42, 2.14) | 1.06 (0.90, 1.26) |
| 2 | $1 039 (881, 1 197) | $549 (369, 729) | 1.93 (1.55, 2.40) | 1.41 (1.18, 1.68) |
| 3–4 | $1 840 (1 587, 2 092) | $1 349 (1 068, 1 631) | 3.32 (2.62, 4.21) | 1.89 (1.58, 2.27) |
| Men§ | ||||
| 0 | $351 (294, 408) | Referent | Referent | Referent |
| 1 | $615 (496, 734) | $264 (135, 393) | 2.15 (1.70, 2.72) | 1.03 (0.86, 1.24) |
| 2 | $932 (727, 1,137) | $581 (364, 797) | 2.45 (1.85, 3.23) | 1.45 (1.17, 1.79) |
| 3–4 | $1 411 (1 125, 1 697) | $1 060 (757, 1 363) | 3.33 (2.50, 4.44) | 1.84 (1.49, 2.28) |
Notes: Costs are in 2023 U.S. dollars. CI = confidence interval; CMS-HCC = Centers for Medicare & Medicaid Services Hierarchical Condition Categories; HHC = home healthcare; SNF = skilled nursing facility.
*Independent variables in models include geographic region, CMS-HCC score, number of functional impairments, and category of frailty phenotype.
†Overall p value for association of functional impairments with incremental inpatient costs <.007 in women and <.002 in men.
‡Overall p value for association of functional impairments with incremental SNF costs <.001 in women and men.
§Overall p value for association of functional impairments with incremental home healthcare costs <.001 in women and men.
After accounting for multimorbidity and functional impairments, mean estimated inpatient expenditures among robust individuals was $4 228 (95% CI: $2 992, $5 465) in women and $6 372 (95% CI: $4 997, $7 746) in men (Table 4). Average incremental inpatient costs were $4 100 (95% CI: $2 269, $5 930) in frail versus robust women and $1 321 (not significantly different from zero, 95% CI: −$33, $2 675) in prefrail versus robust women. Additional inpatient costs among frail women were primarily due to their 2.2-fold greater odds of hospitalization (OR: 2.24, 95% CI: 1.75, 2.86). In contrast, neither prefrailty nor frailty was independently associated with the odds of hospitalization in men; CIs around estimates of incremental inpatient costs were wider and overlapped zero.
Table 4.
Effects of the Frailty Phenotype on Annualized Sector-Specific Costs and Utilization*
| Mean Predicted Costs (95% CI) | Mean Incremental Costs (95% CI) | Odds Ratio of Utilization (95% CI) | Cost Ratio Among Those With Utilization (95% CI) | |
|---|---|---|---|---|
| Inpatient sector | ||||
| Women† | ||||
| Robust | $4 228 (2 992, 5 465) | Referent | Referent | Referent |
| Prefrail | $5 549 (4 840, 6 258) | $1 321 (−33, 2 675) | 1.63 (1.33, 2.01) | 1.04 (0.78, 1.38) |
| Frail | $8 328 (7 213, 9 443) | $4 100 (2 269, 5 930) | 2.24 (1.75, 2.86) | 1.37 (0.99, 1.88) |
| Men† | ||||
| Robust | $6 372 (4 997, 7 746) | Referent | Referent | Referent |
| Prefrail | $7 130 (6 168, 8 092) | $758 (−734, 2 250) | 0.99 (0.84, 1.16) | 1.13 (0.91, 1.40) |
| Frail | $8 085 (6 570, 9 601) | $1 714 (−560, 3 987) | 1.10 (0.86, 1.41) | 1.22 (0.90, 1.64) |
| SNF sector | ||||
| Women‡ | ||||
| Robust | $664 (340, 987) | Referent | Referent | Referent |
| Prefrail | $1 259 (1 047, 1 470) | $595 (238, 952) | 1.96 (1.33, 2.89) | 1.09 (0.76, 1.55) |
| Frail | $2 243 (1 941, 2 546) | $1 579 (1 106, 2 053) | 3.11 (2.05, 4.72) | 1.37 (0.94, 1.99) |
| Men‡ | ||||
| Robust | $420 (223, 616) | Referent | Referent | Referent |
| Prefrail | $920 (723, 1 117) | $501 (244, 757) | 1.69 (1.19, 2.39) | 1.40 (0.96, 2.03) |
| Frail | $1 673 (1 268, 2 079) | $1 254 (766, 1 741) | 2.36 (1.55, 3.57) | 1.94 (1.26, 2.99) |
| HHC sector | ||||
| Women§ | ||||
| Robust | $577 (401, 753) | Referent | Referent | Referent |
| Prefrail | $772 (673, 871) | $195 (14, 376) | 1.54 (1.18, 2.01) | 1.03 (0.80, 1.32) |
| Frail | $1 222 (1 102, 1 342) | $645 (421, 869) | 2.33 (1.73, 3.15) | 1.31 (1.01, 1.71) |
| Men§ | ||||
| Robust | $355 (260, 451) | Referent | Referent | Referent |
| Prefrail | $573 (492, 655) | $218 (110, 326) | 1.51 (1.19, 1.92) | 1.23 (1.00, 1.51) |
| Frail | $881 (751, 1 012) | $526 (351, 701) | 2.17 (1.60, 2.94) | 1.52 (1.19, 1.95) |
Notes: Costs are in 2023 U.S. dollars. CI = confidence interval; CMS-HCC = Centers for Medicare & Medicaid Services Hierarchical Condition Categories; HHC = home healthcare; SNF = skilled nursing facility.
*Independent variables in models include geographic region, CMS-HCC score, number of functional impairments, and category of frailty phenotype.
†Overall p value for association of the frailty phenotype with incremental inpatient costs <.001 in women and .33 in men.
‡Overall p value for association of the frailty phenotype with incremental SNF costs <.001 in women and men.
§Overall p value for association of the frailty phenotype with incremental home healthcare costs <.001 in women and men.
Functional Impairments, Phenotypic Frailty, and SNF Costs
Mean predicted SNF costs among individuals without impairments were $931 (95% CI: $733, $1 129) in women and $618 (95% CI: $454, $781) in men (Table 3). Average incremental SNF costs incurred by individuals with versus without impairments ranged from $819 (95% CI: $412, $1 227) for 1 impairment to $1 572 (95% CI: $1 010, $2 135) for 3–4 impairments among women, and from $639 (95% CI: $239, $1 040) for 1 impairment to $1 446 (95% CI: $713, $2 179) for 3–4 impairments in men. These additional costs among impaired individuals were due to higher odds of an SNF stay and increased costs among those with a stay. Compared with individuals without impairments, 3–4 impairments in women were associated with 2.6-fold higher odds of an SNF stay (OR: 2.62, 95% CI: 1.97, 3.46) and 1.3-fold higher costs among women with a stay (CR: 1.30, 95% CI: 1.03, 1.64), and 3–4 impairments in men was associated with 2.8-fold higher odds of SNF stay (OR: 2.81, 95% CI: 1.97, 4.01) and 1.4-fold higher costs among men with a stay that did not quite reach significance (CR: 1.41, 95% CI: 0.98, 2.01).
Mean estimated SNF expenditures among robust individuals were $664 (95% CI: $340, $987) in women and $420 (95% CI: $223, $616) in men (Table 4). Average incremental SNF costs versus robust ranged from $595 (95% CI: $238, $952) for prefrailty to $1 579 (95% CI: $1 106, $2 053) for frailty in women, and from $501 (95% CI: $244, $757) for prefrailty to $1 254 (95% CI: $766, $869) for frailty in men. These additional costs in frail individuals were driven by higher odds of an SNF stay and increased costs among those with a stay. Frailty versus robust in women was associated with 3.1-fold higher odds of an SNF stay (OR: 3.11, 95% CI: 2.05, 4.72), and 1.4-fold higher (albeit not quite significant) costs among women with a stay (CR: 1.37, 95% CI: 0.94, 1.99), and frailty versus robust in men was associated with 2.4-fold higher odds of SNF stay (OR: 2.36, 95% CI: 1.55, 3.15) and 1.9-fold higher costs among men with a stay (CR: 1.94, 95% CI: 1.26, 2.99).
Functional Impairments, Phenotypic Frailty, and HHC Costs
Mean predicted HHC costs among individuals without impairments were $490 (95% CI: $411, $569) in women and $351 (95% CI: $294, $408) in men (Table 3). Average incremental HHC costs incurred by individuals with versus without impairments ranged from $250 (95% CI: $114, $386) for 1 impairment to $1 349 (95% CI: $1 068, $1 631) for 3–4 impairments among women, and from $264 (95% CI: $135, $393) for 1 impairment to $1 060 (95% CI: $757, $1 363) for 3–4 impairments among men. These incremental costs among impaired individuals were due to higher odds of HHC use and increased costs among those with HHC utilization. For example, 3–4 impairments in women were associated with 3.3-fold higher odds of HHC use (OR: 3.32, 95% CI: 2.62, 4.21) and 1.9-fold higher costs among women with HHC utilization (CR: 1.89, 95% CI: 1.58, 2.27). Results were nearly identical among men.
Mean predicted HHC expenditures among robust individuals were $577 (95% CI: $401, $753) in women and $351 (95% CI: $260, $451) in men (Table 4). Compared with robust, average incremental HHC costs were $195 (95% CI: $14, $376) in women and $218 (95% CI: $110, $326) in men for prefrailty and $645 (95% CI: $421, $869) in women and $526 (95% CI: $351, $701) in men for frailty. The additional costs in frail individuals were driven by higher odds of HHC use and increased costs among those with HHC utilization. Frailty in women was associated with 2.3-fold higher odds of HHC use (OR: 2.22, 95% CI: 1.73, 3.15) and 1.3-fold higher costs among those with HHC utilization (CR: 1.31, 95% CI: 1.01, 1.71). Findings were similar in men.
Contribution of Sector-Specific Costs to Total Healthcare Costs by Category of Functional Impairments and Frailty Phenotype
In both women and men, increases in mean predicted total healthcare costs adjusted for multimorbidity and the frailty phenotype with a greater number of functional impairments were primarily driven by increases in inpatient costs, with smaller increments in SNF and HHC costs (Figure 1; Supplementary Table 5). Among women, more than half of the increase in mean predicted total healthcare costs adjusted for multimorbidity and functional impairments with a greater degree of frailty were due to increases in inpatient costs, with smaller increments in SNF and HHC costs (Figure 1; Supplementary Table 6). While the pattern was similar in men, the gradient in mean total costs across category of frailty phenotype was smaller in magnitude and not significant. Average adjusted outpatient costs in both sexes were not significantly higher with a greater number of functional impairments or greater degree of frailty (Supplementary Tables 5 and 6).
Figure 1.
Mean annualized predicted total healthcare expenditures and costs in specific healthcare sectors in women and men by number of functional impairments (A) and category of frailty phenotype (B). Independent variables in models include geographic region, CMS-HCC score, functional impairments, and frailty phenotype. CMS-HCC = Centers for Medicare & Medicaid Services Hierarchical Conditions Categories; HHC = home healthcare; SNF = skilled nursing facility.
Sensitivity Analyses
Incremental costs of functional impairments and phenotypic frailty in inpatient, SNF, and HHC sectors were not substantially altered when models were further adjusted for Kim CFI (Supplementary Table 7), ADI (Supplementary Table 8), or cognitive impairment (Supplementary Table 9).
Discussion
We found that community-dwelling Medicare beneficiaries with functional impairments have higher subsequent inpatient, SNF, and HHC costs among both sexes after accounting for multimorbidity and the frailty phenotype. After accounting for multimorbidity and functional impairments, phenotypic frailty is independently associated with higher subsequent inpatient costs in women, and higher SNF and HHC costs in both women and men. Reducing costs in older adults with functional impairments or phenotypic frailty will require effective home- or community-based interventions that prevent or delay progression to disability and overt frailty, thereby lowering risk, duration, or complexity of care delivered in these sectors.
We previously reported substantial incremental total healthcare expenditures attributable to functional impairments in both sexes and phenotypic prefrailty and frailty in women not attributable to claims-based indicators of multimorbidity and frailty or each other (10). However, prior investigations have not quantified independent effects of these 2 geriatric domains on costs in specific healthcare sectors. Disability manifested by dependence in basic ADL has been associated with higher utilization of inpatient, home health, and nursing home services (11,13) and an increase in total expenditures in the year after hospitalization (12). Increased risks of all-cause hospitalization (14), admission to hospital from the emergency department (15), preventable hospitalization (34), and hospital readmission (35) have been reported in older community-dwelling beneficiaries with difficulty performing multiple basic or instrumental ADLs. Although most prior investigations accounted for multimorbidity, they did not consider frailty status, quantify incremental costs in specific sectors, or determine the extent to which increased total expenditures among impaired individuals were driven by sector-specific costs. Our findings regarding associations of phenotypic frailty with risks of hospitalization and SNF stay confirm results of a systematic review examining phenotypic frailty as a predictor of hospitalization among community-dwelling adults (36) and extend results of our previous investigations in single cohorts of older women (16) and men (17). In contrast, we found that neither functional impairments nor phenotypic frailty was independently associated with outpatient costs.
Functional impairments and phenotypic frailty in the current study were independently associated with substantial incremental expenditures in SNF and HHC sectors among both sexes. While functional impairments were associated with higher subsequent inpatient costs in both sexes, phenotypic frailty was independently associated with higher inpatient costs only in women as the higher risk of hospitalization in frail men was explained by their greater burden of multimorbidity and functional impairments. Different findings regarding incremental inpatient expenditures attributable to frailty between sexes may in part reflect the higher frailty prevalence in women (9) and differences in reporting of functional impairments between sexes (37,38).
The substantial additional sector-specific costs of functional impairments and phenotypic frailty in our study were not substantially altered when models were further adjusted for the Kim CFI extending findings from our previous investigation on total healthcare expenditures (10). These results suggest that measurement of self-reported functional impairments and the frailty phenotype depict risk profiles distinct from multimorbidity to a greater degree than a claims-derived frailty indicator based on the deficit accumulation index. While it has been suggested that claims-based frailty indicators may be used to identify a vulnerable population for resource-intensive programs, these indicators are omnibus measures missing crucial information needed to inform selection and targeting of interventions designed to reduce costs, especially among community-dwelling older adults without overt disability and advanced frailty.
Previous research has not delineated whether additional sector-specific costs among impaired or frail individuals are due to a higher likelihood of utilization, increased costs among those with utilization, or both. Substantial incremental inpatient costs of functional impairments in both sexes and phenotypic frailty in women in our study were primarily due to higher risks of hospitalization among women and men with impairments and frail women. In contrast, additional SNF and HHC costs in impaired or frail individuals in both sexes were due to both a higher likelihood of utilization of care in these sectors and increased costs among those with utilization. These findings likely reflect the primary Medicare payment system model during the study time frame that pressured hospitals to reduce costs and instituted penalties for 30-day readmission creating a financial incentive to discharge a patient as soon as possible and spurring a dramatic increase in use of SNF as a discharge destination and HHC after hospital discharge or SNF stay (39–41). It is uncertain whether these relationships will persist with implementation of newer Medicare bundled payment initiatives that include services provided by several providers in multiple sectors (42) and the trend of reduced SNF utilization post-hospital discharge observed during the coronavirus disease 2019 pandemic (43).
Assessment of self-reported function and frailty have largely been ignored by healthcare practitioners, systems, and payers in part due to lack of direct reimbursement and continued emphasis on the traditional medical model. There is also a perception that impairments and frailty are not modifiable. Traditional risk adjustment relies heavily on the importance of multimorbidity in predicting healthcare expenditures, but both the Department of Health and Human Services and CMS have acknowledged the need to incorporate functional status into cost prediction models (44,45). Our brief measure of self-reported function could be completed using secure survey technologies minimizing cost and time required for collection and resources required for implementation. These technologies are increasingly utilized by health systems to collect patient-reported information. However, our findings also suggest that assessing both functional impairments and the frailty phenotype will result in more complete characterization of high-risk individuals. Although we made some modifications to the original frailty phenotype components to enhance practicality of their application, assessment of weakness requires grip strength measurement and assessment of slowness necessitates gait speed measurement. Thus, future research is needed to determine whether frailty assessed using simpler instruments, such as the SOF frailty phenotype (46,47) designed for use in the time-constrained clinical practice setting, is associated with healthcare costs, including sector-specific expenditures.
A pressing need for innovative home- or community-based interventions is highlighted by our finding that neither functional impairments nor phenotypic frailty was associated with higher outpatient costs suggesting the presence of financial barriers or poorer access to outpatient care among impaired or frail older individuals. Importantly, accumulating evidence indicates that home-based multidisciplinary programs with components of nursing, occupational or physical therapy, and simple home repair/environmental modifications reduce disability and may provide savings in total expenditures in community-dwelling older adults with functional impairments or physical frailty (48–50). Additional research is warranted to evaluate whether these programs are effective over longer time periods and determine effects of these programs on sector-specific costs.
Strengths of our study include its prospective design, comprehensive phenotyping of participants using cohort study data, and cross-cohort linkage of participants to their Medicare claims. However, this study has limitations. Our cohort was primarily comprised of non-Hispanic White participants. Additional research is needed to determine whether findings are confirmed in minority populations. Our results do not apply to institutionalized older adults. Medication costs, indirect costs such as caregiving from family or friends, or health product expenditures not at least partially covered by Medicare were not captured. Our analysis was limited to participants enrolled in FFS plans. However, we previously reported that characteristics including functional impairments and phenotypic frailty were similar between FFS and non-FFS participants (10). Socioeconomic status and cognition may be associated with healthcare expenditures in older adults, but results were not substantially altered in analyses further adjusted for an index of neighborhood socioeconomic deprivation or cognitive impairment. Finally, our findings require confirmation using more recent data from a nationally representative cohort of older adults linked with Medicare claims.
In this study of older community-dwelling beneficiaries, functional impairments were associated with higher subsequent expenditures in inpatient, SNF, and HHC sectors among both sexes after accounting for multimorbidity and the frailty phenotype. After accounting for multimorbidity and functional impairments, phenotypic frailty was associated with higher subsequent inpatient costs in women, and higher SNF and HHC costs in both sexes. Assessment of self-reported function and phenotypic frailty identifies high-risk individuals that may benefit from integrative patient-centered home-based interventions aimed at reducing future healthcare expenditures and promoting aging in place.
Supplementary Material
Contributor Information
Kristine E Ensrud, Division of Epidemiology & Community Health, School of Public Health, University of Minnesota, Minneapolis, Minnesota, USA; Department of Medicine, University of Minnesota, Minneapolis, Minnesota, USA.
John T Schousboe, HealthPartners Institute, Rheumatology, Bloomington, Minnesota, USA; Divison of Health Policy & Management, School of Public Health, University of Minnesota, Minneapolis, Minnesota, USA.
Allyson M Kats, Division of Epidemiology & Community Health, School of Public Health, University of Minnesota, Minneapolis, Minnesota, USA.
Brent C Taylor, Department of Medicine, University of Minnesota, Minneapolis, Minnesota, USA; Center for Care Delivery & Outcomes Research, VA Health Care System, Minneapolis, Minnesota, USA.
Wei Duan-Porter, Department of Medicine, University of Minnesota, Minneapolis, Minnesota, USA; Center for Care Delivery & Outcomes Research, VA Health Care System, Minneapolis, Minnesota, USA.
Kerry M Sheets, Department of Medicine, University of Minnesota, Minneapolis, Minnesota, USA; Department of Medicine, Hennepin Healthcare, Minneapolis, Minnesota, USA.
Cynthia M Boyd, Departments of Medicine, Health Policy & Management, and Epidemiology, Johns Hopkins University, Baltimore, Maryland, USA.
Peggy M Cawthon, California Pacific Medical Center Research Institute, San Francisco, California, USA.
Lisa Langsetmo, Department of Medicine, University of Minnesota, Minneapolis, Minnesota, USA; HealthPartners Institute, Rheumatology, Bloomington, Minnesota, USA.
Funding
This project is supported by the National Institute on Aging (NIA) under grant number R01 AG067973. The Study of Osteoporotic Fractures was supported by National Institutes of Health (NIH) funding. The NIA provided support under the following grant numbers: R01 AG005407, R01 AR35582, R01 AR35583, R01 AR35584, R01 AG005394, R01 AG027574, and R01 AG027576. The Osteoporotic Fractures in Men Study is supported by NIH funding. The following institutes provide support: the NIA, the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the National Center for Advancing Translational Sciences, and NIH Roadmap for Medical Research under the following grant numbers: R01 AG066671, U01 AG027810, U01 AG042124, U01 AG042139, U01 AG042140, U01 AG042143, U01 AG042145, U01 AG042168, U01 AR066160, and UL1 TR002369. The Health, Aging and Body Composition study was supported by NIA contracts #N01-AG-6-2101; N01-AG-6-2103; N01-AG-6-2106; NIA grant R01-AG028050; NINR grant R01-NR012459. The National Health and Aging Trends Study is supported by the NIA under grant number U01 AG32947.
Conflict of Interest
C.M.B. reports honoraria from UpToDate and DynaMed. The other authors disclose no conflict.
Data Availability
SOF data are available to the public via the “SOF Online” website (https://sofonline.ucsf.edu/). MrOS data are available to the public via the “MrOS Online” website (https://mrosonline.ucsf.edu/). Health ABC data are available to the public via the “Health ABC Study” website (https://healthabc.nia.nih.gov/). NHATS data are available to the public via the “National Health & Aging Trends Study” website (https://www.nhats.org/). Medicare claims data for participants in SOF, MrOS, Health ABC, and NHATS studies are not publicly available. The Centers for Medicare & Medicaid Services (CMS) enters into Data Use Agreements with data requesters for disclosures of protected health information and/or personally identifiable information to ensure that data requesters adhere to CMS privacy and security requirements and data release policies. Requests for statistical code used in this analysis may be directed to L.L. at langs005@umn.edu.
Disclaimer
This is the result of work supported with resources and use of facilities of the Minneapolis VA Health Care System. The views expressed here are those of the authors and do not necessarily reflect the position or policy of the U.S. Department of Veterans Affairs or the U.S. Government.
References
- 1. Lassman D, Hartman M, Washington B, Andrews K, Catlin A.. US health spending trends by age and gender: selected years 2002–10. Health Aff (Millwood). 2014;33(5):815–822. 10.1377/hlthaff.2013.1224 [DOI] [PubMed] [Google Scholar]
- 2. Congressional Budget Office. High-Cost Medicare Beneficiaries. May 1, 2005. Accessed August 1, 2024. www.cbo.gov/publication/16487 [Google Scholar]
- 3. Long P, Abrams M, Milstein A, et al. Effective Care for High-Need Patients: Opportunities for Improving Outcomes, Value, and Health. National Academy of Medicine; 2017. Accessed August 1, 2024. https://nam.edu/wp-content/uploads/2017/06/Effective-Care-for-High-Need-Patients.pdf [PubMed] [Google Scholar]
- 4. Perkins AJ, Kroenke K, Unützer J, et al. Common comorbidity scales were similar in their ability to predict health care costs and mortality. J Clin Epidemiol. 2004;57(10):1040–1048. 10.1016/j.jclinepi.2004.03.002 [DOI] [PubMed] [Google Scholar]
- 5. Centers for Medicare & Medicaid Services. Risk Adjustment. May 27, 2022. Accessed June 11, 2024. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors [PubMed] [Google Scholar]
- 6. Centers for Medicare & Medicaid Services. Report to Congress. December 2021. Accessed June 11, 2024. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors-Items/ReportToCongress [Google Scholar]
- 7. Berkman ND, Chang E, Seibert J, Ali R.. Characteristics of high-need, high-cost patients: a “Best-Fit” framework synthesis. Ann Intern Med. 2022;175(12):1728–1741. 10.7326/M21-4562 [DOI] [PubMed] [Google Scholar]
- 8. Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G.. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci. 2004;59(3):255–263. 10.1093/gerona/59.3.m255 [DOI] [PubMed] [Google Scholar]
- 9. Hoogendijk EO, Afilalo J, Ensrud KE, Kowal P, Onder G, Fried LP.. Frailty: implications for clinical practice and public health. Lancet. 2019;394(10206):1365–1375. 10.1016/S0140-6736(19)31786-6 [DOI] [PubMed] [Google Scholar]
- 10. Ensrud KE, Schousboe JT, Kats AM, Taylor BC, Boyd CM, Langsetmo L.. Incremental health care costs of self-reported functional impairments and phenotypic frailty in community-dwelling older adults: a prospective cohort study. Ann Intern Med. 2023;176(4):463–471. 10.7326/M22-2626 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Fried TR, Bradley EH, Williams CS, Tinetti ME.. Functional disability and health care expenditures for older persons. Arch Intern Med. 2001;161(21):2602–2607. 10.1001/archinte.161.21.2602 [DOI] [PubMed] [Google Scholar]
- 12. Greysen SR, Stijacic Cenzer I, Boscardin WJ, Covinsky KE.. Functional impairment: an unmeasured marker of Medicare costs for postacute care of older adults. J Am Geriatr Soc. 2017;65(9):1996–2002. 10.1111/jgs.14955 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. McColl MA, Shortt S, Gignac M, Lam M.. Disentangling the effects of disability and age on health service utilisation. Disabil Rehabil. 2011;33(13–14):1253–1261. 10.3109/09638288.2010.526166 [DOI] [PubMed] [Google Scholar]
- 14. Perrin NA, Stiefel M, Mosen DM, Bauck A, Shuster E, Dirks EM.. Self-reported health and functional status information improves prediction of inpatient admissions and costs. Am J Manag Care. 2011;17(12):e472–e478. https://www.ajmc.com/view/self-reported-health-and-functional-status-information-improves-prediction-of-inpatient-admissions-and-costs [PubMed] [Google Scholar]
- 15. Smulowitz PB, Weinreb G, McWilliams JM, O’Malley AJ, Landon BE.. Association of functional status, cognition, social support, and geriatric syndrome with admission from the emergency department. JAMA Intern Med. 2023;183(8):784–792. 10.1001/jamainternmed.2023.2149 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Ensrud KE, Kats AM, Schousboe JT, et al. ; Study of Osteoporotic Fractures. Frailty phenotype and healthcare costs and utilization in older women. J Am Geriatr Soc. 2018;66(7):1276–1283. 10.1111/jgs.15381 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Ensrud KE, Kats AM, Schousboe JT, et al. ; Osteoporotic Fractures in Men Study (MrOS). Frailty phenotype and healthcare costs and utilization in older men. J Am Geriatr Soc. 2020;68(9):2034–2042. 10.1111/jgs.16522 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Kim DD, Silver MC, Kunst N, Cohen JT, Ollendorf DA, Neumann PJ.. Perspective and costing in cost-effectiveness analysis, 1974–2018. PharmacoEcon. 2020;38(10):1377–1377. 10.1007/s40273-020-00968-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Sittimart M, Rattanavipapong W, Mirelman AJ, et al. An overview of the perspectives used in health economic evaluations. Cost Eff Resour Alloc. 2024;22(1):41. 10.1186/s12962-024-00552-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Schousboe JT, Paudel ML, Taylor BC, et al. Estimation of standardized hospital costs from Medicare claims that reflect resource requirements for care: impact for cohort studies linked to Medicare claims. Health Serv Res. 2014;49(3):929–949. 10.1111/1475-6773.12151 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Schousboe JT, Paudel ML, Taylor BC, et al. Pre-fracture individual characteristics associated with high total health care costs after hip fracture. Osteoporos Int. 2017;28(3):889–899. 10.1007/s00198-016-3803-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Schousboe JT, Paudel ML, Taylor BC, et al. Estimating true resource costs of outpatient care for Medicare beneficiaries: standardized costs versus Medicare payments and charges. Health Serv Res. 2016;51(1):205–219. 10.1111/1475-6773.12318 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Fried LP, Tangen CM, Walston J, et al. ; Cardiovascular Health Study Collaborative Research Group. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146–M156. 10.1093/gerona/56.3.m146 [DOI] [PubMed] [Google Scholar]
- 24. Alley DE, Shardell MD, Peters KW, et al. Grip strength cutpoints for the identification of clinically relevant weakness. J Gerontol A Biol Sci Med Sci. 2014;69(5):559–566. 10.1093/gerona/glu011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Abellan van Kan G, Rolland Y, Andrieu S, et al. Gait speed at usual pace as a predictor of adverse outcomes in community-dwelling older people an International Academy on Nutrition and Aging (IANA) Task Force. J Nutr Health Aging. 2009;13(10):881–889. 10.1007/s12603-009-0246-z [DOI] [PubMed] [Google Scholar]
- 26. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: report of the European Working Group on Sarcopenia in Older People. Age Ageing. 2010;39(4):412–423. 10.1093/ageing/afq034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Bandeen-Roche K, Seplaki CL, Huang J, et al. Frailty in older adults: a nationally representative profile in the United States. J Gerontol A Biol Sci Med Sci. 2015;70:1427–1434. 10.1093/gerona/glv133 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Centers for Medicare & Medicaid Services. 2009 CMS-HCC, ESRD and Rx Risk Adjustment Software. 2006–2011 Model Software/ICD-9-CM Mappings. CMS. May 22, 2009. Accessed June 11, 2024. https://www.cms.gov/Medicare/Health-Plans/MedicareAdvtgSpecRateStats/Risk-Adjustors-Items/Risk2006-2011 [Google Scholar]
- 29. Pope GC, Kautter J, Ellis RP, et al. Risk adjustment of Medicare capitation payments using the CMS-HCC model. Health Care Financ Rev. 2004;25(4):119–141. https://www.cms.gov/research-statistics-data-and-systems/research/healthcarefinancingreview/list-of-past-articles-items/cms1191786 [PMC free article] [PubMed] [Google Scholar]
- 30. Kim DH, Schneeweiss S, Glynn RJ, Lipsitz LA, Rockwood K, Avorn J.. Measuring frailty in Medicare data: development and validation of a claims-based frailty index. J Gerontol A Biol Sci Med Sci. 2018;73(7):980–987. 10.1093/gerona/glx229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Kind AJH, Buckingham WR.. Making neighborhood-disadvantage metrics accessible—the Neighborhood Atlas. N Engl J Med. 2018;378(26):2456–2458. 10.1056/NEJMp1802313 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. U.S. Department of Commerce. Census Regions and Divisons of the United States. April 17, 2013. Accessed June 11, 2024. https://www2.census.gov/geo/pdfs/maps-data/maps/reference/us_regdiv.pdf [Google Scholar]
- 33. McCullagh P, Nelder JA.. Generalized Linear Models. 2nd ed. Routledge; 1989. 10.1201/9780203753736 [DOI] [Google Scholar]
- 34. Johnston KJ, Wen H, Schootman M, Joynt Maddox KE.. Association of patient social, cognitive, and functional risk factors with preventable hospitalizations: implications for physician value-based payment. J Gen Intern Med. 2019;34(8):1645–1652. 10.1007/s11606-019-05009-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Greysen SR, Stijacic Cenzer I, Auerbach AD, Covinsky KE.. Functional impairment and hospital readmission in Medicare seniors. JAMA Intern Med. 2015;175(4):559–565. 10.1001/jamainternmed.2014.7756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Kojima G. Frailty as a predictor of hospitalisation among community-dwelling older people: a systematic review and meta-analysis. J Epidemiol Community Health. 2016;70(7):722–729. 10.1136/jech-2015-206978 [DOI] [PubMed] [Google Scholar]
- 37. Jindai K, Nielson CM, Vorderstrasse BA, Quiñones AR.. Multimorbidity and functional limitations among adults 65 or older, NHANES 2005–2012. Prev Chronic Dis. 2016;13:E151. 10.5888/pcd13.160174 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Zunzunegui MV, Alvarado BE, Guerra R, Gómez JF, Ylli A, Guralnik JM.. The mobility gap between older men and women: the embodiment of gender. Arch Gerontol Geriatr. 2015;61(2):140–148. 10.1016/j.archger.2015.06.005 [DOI] [PubMed] [Google Scholar]
- 39. Burke RE, Juarez-Colunga E, Levy C, Prochazka AV, Coleman EA, Ginde AA.. Rise of post-acute care facilities as a discharge destination of US hospitalizations. JAMA Intern Med. 2015;175(2):295–296. 10.1001/jamainternmed.2014.6383 [DOI] [PubMed] [Google Scholar]
- 40. Mechanic R. Post-acute care--the next frontier for controlling Medicare spending. N Engl J Med. 2014;370(8):692–694. 10.1056/nejmp1315607 [DOI] [PubMed] [Google Scholar]
- 41. Werner RM, Konetzka RT.. Trends in post-acute care use among Medicare beneficiaries: 2000 to 2015. JAMA. 2018;319(15):1616–1617. 10.1001/jama.2018.2408 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Yee CA, Pizer SD, Frakt A.. Medicare’s bundled payment initiatives for hospital-initiated episodes: evidence and evolution. Milbank Q. 2020;98(3):908–974. 10.1111/1468-0009.12465 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Werner RM, Bressman E.. Trends in post-acute care utilization during the COVID-19 pandemic. J Am Med Dir Assoc. 2021;22(12):2496–2499. 10.1016/j.jamda.2021.09.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Medicaid. Managed Long-Term Services and Supports. Accessed August 1, 2024. https://www.medicaid.gov/medicaid/managed-care/managed-long-term-services-and-supports/index.html [Google Scholar]
- 45. National Quality Forum (NQF). NQF: Risk Adjustment Technical Guidance Final Report—Phase 2. Accessed August 1, 2024. https://www.qualityforum.org/Publications/2022/12/Risk_Adjustment_Technical_Guidance_Final_Report_-_Phase_2.aspx [Google Scholar]
- 46. Ensrud KE, Ewing SK, Taylor BC, et al. Comparison of 2 frailty indexes for prediction of falls, disability, fractures, and death in older women. Arch Intern Med. 2008;168(4):382–389. 10.1001/archinternmed.2007.113 [DOI] [PubMed] [Google Scholar]
- 47. Ensrud KE, Ewing SK, Cawthon PM, et al. ; Osteoporotic Fractures in Men Research Group. A comparison of frailty indexes for the prediction of falls, disability, fractures, and mortality in older men. J Am Geriatr Soc. 2009;57(3):492–498. 10.1111/j.1532-5415.2009.02137.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48. Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Byers A.. A program to prevent functional decline in physically frail, elderly persons who live at home. N Engl J Med. 2002;347(14):1068–1074. 10.1056/NEJMoa020423 [DOI] [PubMed] [Google Scholar]
- 49. Szanton SL, Xue QL, Leff B, et al. Effect of a biobehavioral environmental approach on disability among low-income older adults: a randomized clinical trial. JAMA Intern Med. 2019;179(2):204–211. 10.1001/jamainternmed.2018.6026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Szanton SL, Leff B, Li Q, et al. CAPABLE program improves disability in multiple randomized trials. J Am Geriatr Soc. 2021;69(12):3631–3640. 10.1111/jgs.17383 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
SOF data are available to the public via the “SOF Online” website (https://sofonline.ucsf.edu/). MrOS data are available to the public via the “MrOS Online” website (https://mrosonline.ucsf.edu/). Health ABC data are available to the public via the “Health ABC Study” website (https://healthabc.nia.nih.gov/). NHATS data are available to the public via the “National Health & Aging Trends Study” website (https://www.nhats.org/). Medicare claims data for participants in SOF, MrOS, Health ABC, and NHATS studies are not publicly available. The Centers for Medicare & Medicaid Services (CMS) enters into Data Use Agreements with data requesters for disclosures of protected health information and/or personally identifiable information to ensure that data requesters adhere to CMS privacy and security requirements and data release policies. Requests for statistical code used in this analysis may be directed to L.L. at langs005@umn.edu.