SCIENTIFIC INTEGRITY & QUALITY IN LONG-TERM CARE RESEARCH: A RESPONSE TO THE ASPE STUDY
By: David E. Kingsley, PhD
INTRODUCTION
In an open and just society, research that meaningfully shapes public policy must meet the highest ethical and scientific standards. When slipshod or methodologically unsound work threatens to distort public understanding or undermine social and economic justice, trained and experienced professionals have an obligation to challenge its use in legislation and regulation.
A recent report sponsored by the HHS Assistant Secretary for Planning and Evaluation (ASPE)—Medicaid Payment Rates and Costs of Caring for the Medicaid Population Residing in Nursing Homes—falls far short of these standards.[i] The study is seriously flawed in design, execution, and interpretation, and warrants rigorous scrutiny. As published, it risks reinforcing a false industry narrative that for‑profit nursing home chains are chronically under‑compensated, a claim that is not supported by the underlying data.
Framing and falsehoods that underpin the industry’s long‑standing hardship narrative have real consequences: they are routinely invoked to justify substandard care, weak oversight, and continued financial opacity. The ASPE report, if left unchallenged, risks lending official legitimacy to this narrative by asserting that Medicaid reimburses only 82 percent of providers’ costs.
That conclusion rests on a poorly designed statistical study and misunderstanding of the role of facility cost reports in corporate finance and accounting. As this article will demonstrate, the regression models used to generate the 82 percent figure are methodologically invalid, and the very construct of “cost” entered as the dependent variable is conceptually unsound. The authors’ failure to correctly conceptualize basic finance and accounting principles renders the study fatally flawed.
Accordingly, this paper will show that:
Nursing home cost reports are inadequate for evaluating financial returns or true economic performance;
Medicaid payments are routinely increased through related‑party transactions and financial engineering;
Actual costs and cash flows are obscured by a deliberately opaque network of shell companies and subsidiaries through which Medicaid revenue is routed andamplified;
Uni-level regression models that reduce a complex, dynamic, multi-level social system to a few coefficients, lack statistical validity.
“Costs,” as a critical measure based on line 89, column 7 of Form A-1, lacks construct validity in the creation of a dependent (outcome) variable despite post hoc adjustments.
SECTION I: CASH FLOW & THE COST CONSTRUCT IN NURSING HOME RESEARCH
Cash is King – Operating Net/Margin is Not a Measure of Financial Loss or Gain
In contemporary corporate finance and accounting, cash flow[i]—not “margin,” “profit,” or similar terms—is the accepted standard for measuring economic returns. Cash flow reflects the actual movement of money into and out of a company and therefore provides the only reliable basis for evaluating financial outcomes for investors. Conversely, nursing home cost reports are a poor proxy for return on investment and offer little insight into the economic performance of facilities or parent/holding companies. Treating a dollar of Medicaid revenue as if it were a dollar of cash flow reflects a fundamental misunderstanding of finance and accounting. This distinction is not academic; it is foundational to any empirical study that claims to measure financial adequacy or shortfall.
Optimization of shareholder value is the is the guiding ethos – indeed the fiduciary responsibility – of management philosophy in the current state of U.S. capitalism.[ii] In modern corporate finance, increases in cash distributed to investors, used for expansion or acquisition, placed on the balance sheet, or otherwise reinvested are the sine qua non of successful management. For this reason, the Securities and Exchange Commission requires that every publicly traded company include a cash flow statement in its annual 10‑K filing so that investors can assess the firm’s true financial condition. Nonpublic, closely held companies have no requirement to reveal their consolidated financial statements.
[i] An extensive discussion of cash flow is beyond the scope of this paper. Nevertheless, it must be noted that the income statement of the cost reports (Form G-3) does not inform public policy as it pertains to fair economic returns for profit seeking providers. Net operating income – net patient revenue minus expenses – is often misinterpreted as profit or the ultimate return on investment.
[ii] In business management theory as it has evolved since the 1970s, managers are “agents” of shareholders and have no ethical responsibility for stakeholders such as customers, patients, citizens, and communities. For two excellent discussion of the transition from a 1950s-60s philosophy of corporate public responsibility to today’s ethic of shareholders as the sole beneficiaries of the corporation see: Nicholas Leman (2019) Transaction Man: The Rise of the Deal and the Decline of the American Dream, New York: Farrar, Straus, & Giroux; Nicholas Shaxson (2019) Finance Curse: How Global Finance is Making Us All Poorer, New York: Grove Press.
When researchers substitute reported “costs” for cash flow, they collapse the distinction between accounting categories and economic returns, rendering any subsequent analysis conceptually unsound. This conceptual error becomes even more consequential when applied to the nursing home sector, where cost reports are not designed to measure economic returns and cannot be repurposed to do that. The “cost” construct used in research at issue in this paper collapses operating expenses, related‑party payments, depreciation schedules, and imputed costs into a single accounting artifact that bears no relationship to the cash actually generated by a facility. As a result, research which relies on cost reports to infer financial performance are not measuring returns, adequacy, or shortfall; they are measuring the internal logic of the cost‑reporting system itself. The distinction matters because the gap between reported “costs” and real cash flow is not a subtle discrepancy—it is a structural divergence that becomes unmistakable when examining an actual facility’s financial footprint.
The clearest way to see this structural mismatch is to walk through a real nursing home cost report, where the numbers themselves reveal how far the cost construct diverges from the economic reality of the enterprise. Section II below, is a walkthrough of several 2004 cost report forms submitted by the largest nursing home corporation in the U.S. These forms are revealing: they show how the cost construct obscures rather than illuminates the economic reality of the enterprise.
SECTION II: FORMS G-3, A-1, & A-5: A COST REPORT WALK THROUGH AND ILLUSTRATION WITH A TYPICAL FACILITY
To illustrate how the cost‑report construct obscures rather than clarifies financial performance, we turn to a concrete example: the Big Blue/Riverbend facility owned by the Ensign Group. This facility’s cost report provides a textbook demonstration of the structural gap between reported “costs” and actual economic returns. On paper, the report presents a coherent accounting narrative; in practice, it embeds depreciation schedules, related‑party payments, and imputed capital charges that bear little resemblance to the cash the enterprise generates. By examining the report line by line, we can see how the cost construct inflates capital costs, masks ownership arrangements, and ultimately produces a financial portrait that is analytically incompatible with any measure of return on investment. This walk‑through is not an isolated curiosity—it is a window into the systemic misalignment at the heart of federal cost‑based research.
What makes the Riverbend example especially revealing is that its inflated capital costs do not arise from a single line item but from the interaction of several structural features embedded in the cost‑reporting system. Depreciation is calculated not on the facility’s real asset value but on historical or related‑party book values that can be many times higher than the property’s assessed worth.[i] Lease payments to related real‑estate entities are reported as operating costs even when they function as internal transfers within a vertically integrated ownership chain. Interest expenses may reflect internal loans rather than market‑based financing, allowing owners to generate “costs” by lending money to themselves. Each of these mechanisms is permissible within the reporting framework, yet none of them reflects an actual cash outflow that corresponds to the economic burden implied by the cost report. Together, they create the appearance of a capital‑intensive enterprise where, in reality, the financial structure is engineered to manufacture reportable costs rather than document genuine investment.
[i] The Wyandotte County property database indicates that the building has an assessed valuation or $4 million. A shell company formed as an LLC under Nevada law by the Ensign Group is the listed owner. All of Ensign’s facilities and shell companies are formed as LLCs in Nevada, which, like Delaware is a state with generous tax advantages, high opacity, low liability.
Total costs/expenses reported on line 89 of form A-1 and on the income statement are often misinterpreted as actual expenses, when many non-cash expenses such as depreciation and interest lower taxable income and increase cash distribution to investors. As mentioned earlier, these tax expenditures add value to Medicaid dollars as they flow through the nursing home system.
Every nursing home, every nursing home chain, and every state nursing home system is unique. Nevertheless, the patterns and practices illustrated in the case of Riverbend are common. For instance, line 1, column 5 – capital costs for land and buildings – appears to be highly inflated given the assessed value of the building.
This is an OPCO/PROPCO structured facility. The land and building are owned by an Ensign Group shell company. It is also an OPCO/MANCO structure. The licensee is a shell company through which home office allocations flow. In addition, the facility includes an OPCO/RELPCO component with many of its goods and services purchased from Ensign Group subsidiaries. All of the company’s 400 facilities are LLCs formed in the state of Nevada, which lowers its tax burden and facilitates secrecy.
Table 1 is the 2024 income statement (Form G-3) submitted by the Ensign Group to the Kansas Department of Aging & Disability Services and finalized by the Kansas state auditors. The statement notes net patient revenue (Medicare, Medicaid, and self-pay) of $14,356,777.
Expenses in Table 1 are shown as $13,927,639 – less than net patient revenue even though these expenses include noncash expenses such as depreciation and interest plus payments to related parties, and the home office. So, a large portion of the noted expenses are tax write downs and transfers to related parties at a questionable cost that enhance cash flow. Even with costs that are non-costs folded into expenses shown on the income statement, the facility had a net operating income of 3%.
Table 1:Ensign Group 2024 Form G-3 Income Statement
Tables 2 and 3 are partial views of capital costs (costs for land and buildings) submitted on Form A-1. Line 1, column 5 in Table 3 is the reconciliation with company ledgers for capital costs (land and buildings). It is puzzling that the company’s submitted costs were increased by state auditors by $246,123 as allowable costs (Line 1, column 6). The building is assessed at slightly more than $4 million. Even 25% of a building’s assessed value submitted as capital costs in one year is excessive, but the state increased it to 30%.
A state auditor contacted for an explanation of this extraordinary expense was also perplexed by the state’s action. However, the auditor could not identify a person in the agency who could explain it. This treatment of capital costs for real estate is not uncommon. There is no way of knowing if the costs are due to a triple net lease, plus an interest only loan and accelerated depreciation.
The ASPE report failed to recognize the highly questionable capital costs, home office allocations, and related party payments in their dependent variable construct. Reviews of thousands of facility cost reports suggest that overpriced services and goods are common but ignored by policy makers and researchers.
Table 2: Ensign Group 2024 Form A-1, Lines 1-3, Columns 1-3
Table 3:Ensign Group 2024 Form A-1, Lines 1-3, Columns 4-7
A considerable amount of Medicaid funding flows out of privatized facilities into parent companies through subsidiaries providing goods and services. These subsidiaries are generally limited liability companies. Many are shell companies with no buildings and no employees. Through transfer pricing – pricing higher than fair market value – Medicaid funds are leveraged into increased cash flow.
Tables 4, 5, and 6 show data from the Ensign Group’s submittal of Form A-5, which lists payments to related parties. The company claimed $1,383,636 in payments to their home office and variety of subsidiaries. The Kansas auditor increased that amount by $634,832, which consequently increased allowable costs from related parties to $2,018,568. The public has no way to know the rationale for these decisions by auditors.
The opacity of state auditing and the inaccessibility of government employees making decisions about allowable costs violates the right of the American people to know how their taxes are managed. Secrecy and lack of transparency create conditions for corruption and industry propaganda. The taxpaying public should be able to discern if providers are extracting excessive cash at the expense of patients’ care.
Without transparency, objective, scientific studies cannot be undertaken. Testable hypotheses are only possible in environments where verifiable data is available. The pricing of products and services by subsidiaries of parent/holding companies to its licensed LTC/SKN facility(s) should be compatible with markets. But knowing if that is the case is practically impossible.
Governmental entities and legislators could be considered the main culprits in the opaque, confusing, and in many cases, corrupt practices common in the U.S. nursing home industry. In contemporary business management philosophy, managers are driven by shareholder-return as their guiding ethic. It is not surprising that companies will ignore stakeholder rights and ethical business practices. Hence, the responsibility for honest, ethical management practices rests with regulatory agencies and legislators. They are failing badly.
Table 4:Ensign Group 2024 Form A-5 (Related Parties)
Table 5: Ensign Group 2024 Form A-5 (Related Parties, Columns 2 & 3)
Table 6 Ensign Group 2024 Form A-5 (Related Parties, columns 4–6)
Expenses in the calculation of net operating income include non-cash expenses such as depreciation and interest. Furthermore, transfer pricing by related parties is frequently inflated above market prices. For instance, financiers understand that food purchased in bulk generates mark downs for parent/holding companies (i.e., home offices) but can be priced higher for sale to facility subsidiaries. The same financial machination applies to labor, housekeeping, maintenance, insurance, and other ancillary services.
From a cash flow perspective, an increase in cash flow, and consequently in economic reward for investors, is embedded in costs – otherwise known as expenses. An increase in some highly significant expenses paradoxically increases return on investment, cash and cash equivalents on the balance sheet, and consequently enhances liquidity.
In nursing home finance, cash flow refers to adding value to Medicaid, Medicare, and self-pay funds through financial engineering. In order to enhance cash flow, investors leverage other forms of tax expenditures. For instance, federal and state tax codes allow for write downs for depreciation and interest.
By engaging in tax and lending arbitrage, taxable income can be greatly reduced, thereby increasing availability of cash for distribution to investors, financing acquisitions and other investments, maintenance of liquidity, stock buybacks, and so forth.
Suffice it to say that Medicaid dollars that flow into the privatized nursing home system increase in value through financial engineering not apparent at the facility level. It is practically impossible to measure capital, nay cash, that flows through an opaque network of shell companies with hidden ownership.
Nursing home care involves matters of life and death. Resources required to insure optimal care often depend on political decisions, which in turn depend upon a technically persuasive argument by advocates. Regression models are often applied to data flowing from complex, dynamic social systems with many interacting subsystems. The output produced by regression algorithms is typically intimidating to professionals untrained in mathematical/statistical fundamentals of intermediate or advanced statistics.
Consequently, the results of even extremely poor modeling designs can silence critique. Parties with a vested interest in an outcome can bully adversaries and undermine open and fair discourse. It is the duty of professionals trained and experienced in methods and statistics to challenge unsuitable research designs harmful to the health and well-being of patients. But too often invalid and misleading modeling goes unchallenged.
ASPE Report Statistical Modeling
The authors presented a multiple ordinary least squares regression model. This report will discuss the following inherent and often interrelated errors and fallacies in the model as specified:
Specification Error: Specification error pertains to the selection of independent and dependent variables entered in a linear model. If unrelated variables are selected as predictors of an output variable (e.g. facility expenses/costs), or if a dependent/output variable is not the most valid measure for addressing the issue studied, then results will be misleading and invalid.
Measurement Error: In a multiple regression model, the output/dependent variable is continuous. If nursing home costs are mismeasured, the model as specified and the results of the analysis will be misleading. Valid measure of costs as reflected by facility cost reports is the sine qua non of the model specified by the authors. For instance, capital costs reported on cost reports such as rent, depreciation, and interest should reflect the fair market value of the real property. Often it does not. Therefore, costs are inflated as are a myriad of ancillary services and goods provided by related parties.
Construct Validity: The misunderstanding and selection of costs from line 89, Column 7 of form A-1, has been a theme throughout this report. In addition to the problem with the cost construct, there appears to no rational explanation for the selection of the variables entered into models as independent/predictor variables.
The “Cost” construct apparently refers only to costs reported on facility cost reports. This ignores the OPCO/PROPCO/MANCO structure of for-profit nursing homes. Costs(expenses) reported in facility cost reports actually enhance cash flow – often through LLCs set up as shell companies. For example, Medicaid dollars are leveraged into increased cash flow through write downs for reported capital costs.
All six independent/predictor variables entered into the model were categorical. Four of the six predictor variables were derived from categorization of continuous variables. Number of beds, percentage occupancy, percentage of Medicaid occupancy, and HPRD. The categorization of normally distributed, ratio-level measures is puzzling. These transformations increase error and lower the power of the analysis to accurately and precisely partition variance.
The authors don’t explain how they arrived at their decisions to reduce continuous, ratio level measures to a few categories. For example, they categorized bed size from 0-60, 61-120, 121-180, and 180+. Like the other categorized continuous measures – occupancy rate, Medicaid payer mix, and HPRD – statistical fundamentals suggest that these variables are more suitable for regression analysis as ratio level measures without any transformation. A large amount of error is introduced into the model through categorization. As demonstrated with Model 1 and Model 2 on page 11 below, two nearly identical facilities have meaningfully different predicted values.
Model Fit: Properly presented regression analysis results would include an analysis of residuals with appropriate statistics and graphics. For instance, there is no indication the authors tested their model for autocorrelation, multicollinearity, and overall fit. They did include an r squared for their three models but given the poor design of the model, it is unlikely that those measures are meaningful. That would be particularly the case due to nesting in states discussed below.
The Simple & Multiple OLS Regression Model Explained
The regression models presented by the authors are most certainly intimidating for lay people without specialized training in intermediate or advanced statistics. Nevertheless, they simply state that variables such as Medicaid payment rates, Medicaid costs, and Medicaid payment-to-cost ratios can be predicted by computer algorithms (mathematical formulas) that calculate the effects of a group of predictor variables. For instance, the variables “ownership type,” “Number of beds,” “Occupancy rate,” “Medicaid payer mix,” “HPRD,” and “star rating” were each weighted based on their effects on the outcome variables.
The ratio between the modeled payment rate to modeled cost was entered as a dependent or outcome variable in a third model. These models are highly questionable for a variety of errors in design. The modeling of costs through specifying “costs” as the outcome/predictor variable when the stated costs on Form A-1, line 89, column 7, as already discussed, are nominal costs but not actual costs. To reiterate, these metrics hide cost reductions that add value to Medicaid.
For professionals with an interest in regression analysis, the multiple regression models discussed are presented below symbolically and with actual results showing the problematic nature of categorized variables such as HPRD, bed size, etc.
The purpose of regression analysis is to explain variance. All cases in an analysis can be expected to vary from the average. When other variables are taken into account (controlled for), different levels of the variable of interest reflect higher or lower averages. The average blood pressure of all U.S. residents may be 120 over 85. However, when controlling for age, the average blood pressure will increase linearly by age. Variance decreases at each level of age. Controlling additional variables will decrease variance even more.
Theoretically, adding additional variables will increase the amount of variance explained and allow for prediction of outcome variables such as “reimbursement rate.” However, specifying the correct variables and all of the interaction between variables in a complex, dynamic social system is not only daunting, but also practically impossible. The fallacy that results is known as reductionism. Reducing a social system as complex as the U.S. nursing home industry to a few variables is illogical.
Regression analysis can baffle lay people and add a patina of science to what is essentially an unscientific study. The following is a technical explanation of modeling with ordinary least squares regression modeling for individuals who would like to work through it. The error rate represented by in the models below is an indication of how well a model predicts the outcome variable, e.g., reimbursement rate or cost. That is a critical diagnostic in all appropriately reported regression models but was ignored in the ASPE report.
Adequately presented regression models include an analysis of error. The distance of an actual value from its predicted value is known as a “residual.” Residuals should be analyzed and graphed. An in-depth discussion of regression diagnostics is beyond the scope of this paper. The point is that a paper with a model of grave implications for the nursing home industry and patients should reflect serious scientific and mathematical standards. The ASPE paper reveals a cavalier attitude toward the power and influence of mathematical models.
A Formal Presentation of Regression Analysis
The Model Specified by the Authors: Problematic Categorization of ratio level measure
ASPE researchers for some unexplained reason categorized equal interval/ratio level measures. They presented no evidence that such a transformation is necessary. In fact, variables such HPRD, bed size are normally distributed and in accord with regression assumptions. For instance, Table 7 and the histogram in Figure 2, indicate that HPRD is normally distributed.
By categorizing continuous variables suitable for entry into a regression model, the researchers increased error in their model and undermined the reliability and validity of predicted rates. The following models are application of the ASPE regression coefficients to two different facilities with slightly different demographics but a significant difference in what the model predicts.
One analysis models a for profit facility with 61 beds, 70 percent occupancy, an HPRD of 3.5, 70 percent Medicaid, and 2 stars. The other models a for profit facility with 60 beds, and 3.4 HPRD. All other variables are the same for each facility.
Although the two facilities are, for all practical purposes, exactly the same, the predicted rates of 186.24 and 183.69 are disturbingly different.
Model I: For profit with 61 Beds, 70 percent occupancy, 70 percent Medicaid, 3.5 HPRD, & 2 stars
Model 2: Same as Model 1 except with 60 Beds & 3.4 HPRD
Categorization of continuous variable is puzzling given the suitability of HPRD and the other equal interval, ratio level measurements for regression analysis. Table 7 and the histogram in Figure 1.
Table 7:2024 HPRD Descriptive Statistics
Context Variables & Hierarchical Linear Modeling: The Fallacy of Ignoring Nesting
Ignoring “nesting” is a major fallacy in regression modeling. For instance, if students reading scores across a school district were regressed on family income without accounting for each child’s teacher, and school. Students, as the unit of analysis are nested in teacher/classroom, which is in turn nested in school. Without accounting for the effects of the teacher and the school, the outcome variable – reading score – rests solely on the students’ family income. If reading scores were regressed on family income in one classroom or one school, the intercept and slope in might randomly be the same as in another classroom or school, but the probability of that is quite low.
Nursing homes are nested within states. Every state’s nursing home policy, culture, and demographics is unique. For instance, Kansas has 275 facilities with an average of 60 beds. California’s 1200 facilities average 110 beds. The intercept and regression line would be unique for each state. Therefore, a model that accounts for the variance in state intercepts and trend lines is necessary.[
The two-level model – hierarchical linear model or HLM – would look like this:
Two-Level HLM (Random Intercept) Model
Level-1 (Facility Level)
This models how facilities differ within each state
ASPE Cost Adjustments & The Cost Construct
ASPE designed the critically important dependent variable by making four adjustments to the amount of “Net Expense for Allocation” noted on Form A-1 (line 89, column 7). In the first instance, line 89, column 7 is conceptually a poor measure of costs. This particular notation on the form A-1 is not a measure of “cost of caring for Medicaid residents.” It is a post-reclassification, post-adjustment number that includes:
Direct care costs
Indirect costs
Overhead
Capital-related costs
Ancillary services
Related-party transactions
Reclassified expenses (sometimes aggressively so)
Facility-level allocations that have nothing to with Medicaid patients
The adjustment for therapy is the least conceptually pure measure of “cost” in the analysis. Indeed, on pages 7-8 and in the footnotes, ASPE acknowledges that line 89 includes categories that do not reflect Medicaid cost of care. They adjust costs to account for the fact that Medicare pays more for therapy than Medicaid. This is an ad hoc adjustment, not a cost-based one.
For instance, therapy costs are notoriously inflated due to upcoding. In 2019, some facilities were transitioning to the patient driven payment model (PDPM) to prevent the ease with which facilities could upcode under the resource utilization groups (RUGs) method of weighting rates by RUGs category. While some states were still moving to PDPM in 2019, many were still measuring case mix through the older RUGs methodology. There is no scientific evidence that PDPM is solving the over-billing problem.
Portioning out therapy hours merely by proportion Medicaid versus Medicare without actual therapy hours accurately assigned to one or the other payor ignores the case mix index for both groups. Medicaid CMI as well as Medicare CMI will vary from facility to facility.
They attempt to remove capital costs (depreciation, interest, lease/rent). But capital costs folded into line 89 are: (1) reclassified, (2) mixed with related-party markups – often inflated, and (3) not separable without facility-level detail.
The ASPE researchers attempted to subtract ancillary service costs (lab, radiology, pharmacy, etc.). But ancillary costs are often transfer payments to related parties. There is no evidence to be found in cost reports of the alignment of transfer pricing with market prices. Sometimes these expenses are embedded in overhead. Sometimes they are reclassified into nursing or administrative lines.
They attempted to exclude costs for services not related to nursing home care (e.g., home health, hospice, outpatient therapy). These costs are rare and insignificant in total costs on line 89.
Can costs reported on facility level cost reports be disentangled for the purposes of measuring the value of Medicaid dollars in cash flow? Perhaps they could be if researchers had access to company ledgers and/or owners were transparent regarding property ownership, internal rates of return and cap rates, loans, and so forth. The same kind of transparency would be necessary in regard to the eclectic mix of related party subsidiaries doing business in insurance, therapy, dietary services, housekeeping, and labor contracting.
SUMMARY
Poorly conceptualized and implemented studies of nursing home finance lead to mismanagement of public funds and often justifies subpar care for patients. Indeed, the ASPE report conclusions are erroneous and biased in favor of for-profit investors. Findings from that unscientific study reinforce a false industry narrative of financial hardship imposed on providers through inadequate Medicaid funding.
Most for-profit nursing home companies are closely held. Their consolidated financial statements are not publicly available. So, industry-wide data are not accessible for verification of cash flow and liquidity. Therefore, scientific claims based on facility-level data cannot be verified. Cash flows through a secretive system of shell companies onto the hidden legers of parent/holding companies remain closeted on financial statements closed to the public.
Evidence from the study of a large publicly listed nursing home corporation with a high percentage of Medicaid patients suggests that the business is quite lucrative for investors.[1] Furthermore, it can be deduced from the review of facility cost reports that real estate assets and transfer payments to related parties and home offices support robust cash flow through parent/holding companies to shareholders. This evidence and deductive logic was ignored by the ASPE researchers.
It is remarkable that full costs on line 89, column 7 of Form A-1 was accepted by the researchers at face value and served as the denominator in the study’s critical Medicaid to cost ratio. A post hoc adjustment was made for therapy services, but there is no means by which that can be verified without patient level data and allowance for upcoding. A guesstimate is not a scientific fact.
The invalid, poorly designed, and inappropriately reported regression models will only serve as a barrier to a full and fair communication process. Regression modeling even at the sophomoric level of this study is intimidating to the uninitiated. Coefficients and statistical significance are well understood by professionals who are trained and experienced in the field of statistics and methods.
Common sense should also tell one that a system as complex as the U.S. nursing home system cannot be reduced to a small number of coefficients. However, if data taken from nursing home cost reports are misunderstood and if Medicaid funds are considered to have a rigid and unchanging value as the flow through shell companies and other entities, the results of regression models will be biased in favor of the industry.
[1] D. Kingsley & C. Harrington (2022), “Financial & Quality Metrics of a Large Publicly Traded Nursing Home Chain in the Age of Covid-19.” International Journal of Health Services,1-13.
[1] https://aspe.hhs.gov/reports/assessing-medicaid-payments-costs-nursing-homes
[1] An extensive discussion of cash flow is beyond the scope of this paper. Nevertheless, it must be noted that the income statement of the cost reports (Form G-3) does not inform public policy as it pertains to fair economic returns for profit seeking providers. Net operating income – net patient revenue minus expenses – is often misinterpreted as profit or the ultimate return on investment.
[1] In business management theory as it has evolved since the 1970s, managers are “agents” of shareholders and have no ethical responsibility for stakeholders such as customers, patients, citizens, and communities. For two excellent discussion of the transition from a 1950s-60s philosophy of corporate public responsibility to today’s ethic of shareholders as the sole beneficiaries of the corporation see: Nicholas Leman (2019) Transaction Man: The Rise of the Deal and the Decline of the American Dream, New York: Farrar, Straus, & Giroux; Nicholas Shaxson (2019) Finance Curse: How Global Finance is Making Us All Poorer, New York: Grove Press.
[1] The Wyandotte County property database indicates that the building has an assessed valuation or $4 million. A shell company formed as an LLC under Nevada law by the Ensign Group is the listed owner. All of Ensign’s facilities and shell companies are formed as LLCs in Nevada, which, like Delaware is a state with generous tax advantages, high opacity, low liability.
[1] See: Robert Bickel (2007) Multilevel Analysis for Applied Research. New York: The Guilford Press; Jos W.R. Twisk (2006) Applied Multilevel Analysis. New York: Cambridge University Press;
[1] D. Kingsley & C. Harrington (2022), “Financial & Quality Metrics of a Large Publicly Traded Nursing Home Chain in the Age of Covid-19.” International Journal of Health Services,1-13.
Leave a Comment
Last Updated: February 20, 2026 by David Kingsley, PhD
SCIENTIFIC INTEGRITY & QUALITY IN LONG-TERM CARE RESEARCH: A RESPONSE TO THE ASPE STUDY
By: David E. Kingsley, PhD
INTRODUCTION
In an open and just society, research that meaningfully shapes public policy must meet the highest ethical and scientific standards. When slipshod or methodologically unsound work threatens to distort public understanding or undermine social and economic justice, trained and experienced professionals have an obligation to challenge its use in legislation and regulation.
A recent report sponsored by the HHS Assistant Secretary for Planning and Evaluation (ASPE)—Medicaid Payment Rates and Costs of Caring for the Medicaid Population Residing in Nursing Homes—falls far short of these standards.[i] The study is seriously flawed in design, execution, and interpretation, and warrants rigorous scrutiny. As published, it risks reinforcing a false industry narrative that for‑profit nursing home chains are chronically under‑compensated, a claim that is not supported by the underlying data.
Framing and falsehoods that underpin the industry’s long‑standing hardship narrative have real consequences: they are routinely invoked to justify substandard care, weak oversight, and continued financial opacity. The ASPE report, if left unchallenged, risks lending official legitimacy to this narrative by asserting that Medicaid reimburses only 82 percent of providers’ costs.
That conclusion rests on a poorly designed statistical study and misunderstanding of the role of facility cost reports in corporate finance and accounting. As this article will demonstrate, the regression models used to generate the 82 percent figure are methodologically invalid, and the very construct of “cost” entered as the dependent variable is conceptually unsound. The authors’ failure to correctly conceptualize basic finance and accounting principles renders the study fatally flawed.
Accordingly, this paper will show that:
SECTION I: CASH FLOW & THE COST CONSTRUCT IN NURSING HOME RESEARCH
Cash is King – Operating Net/Margin is Not a Measure of Financial Loss or Gain
In contemporary corporate finance and accounting, cash flow[i]—not “margin,” “profit,” or similar terms—is the accepted standard for measuring economic returns. Cash flow reflects the actual movement of money into and out of a company and therefore provides the only reliable basis for evaluating financial outcomes for investors. Conversely, nursing home cost reports are a poor proxy for return on investment and offer little insight into the economic performance of facilities or parent/holding companies. Treating a dollar of Medicaid revenue as if it were a dollar of cash flow reflects a fundamental misunderstanding of finance and accounting. This distinction is not academic; it is foundational to any empirical study that claims to measure financial adequacy or shortfall.
Optimization of shareholder value is the is the guiding ethos – indeed the fiduciary responsibility – of management philosophy in the current state of U.S. capitalism.[ii] In modern corporate finance, increases in cash distributed to investors, used for expansion or acquisition, placed on the balance sheet, or otherwise reinvested are the sine qua non of successful management. For this reason, the Securities and Exchange Commission requires that every publicly traded company include a cash flow statement in its annual 10‑K filing so that investors can assess the firm’s true financial condition. Nonpublic, closely held companies have no requirement to reveal their consolidated financial statements.
[i] An extensive discussion of cash flow is beyond the scope of this paper. Nevertheless, it must be noted that the income statement of the cost reports (Form G-3) does not inform public policy as it pertains to fair economic returns for profit seeking providers. Net operating income – net patient revenue minus expenses – is often misinterpreted as profit or the ultimate return on investment.
[ii] In business management theory as it has evolved since the 1970s, managers are “agents” of shareholders and have no ethical responsibility for stakeholders such as customers, patients, citizens, and communities. For two excellent discussion of the transition from a 1950s-60s philosophy of corporate public responsibility to today’s ethic of shareholders as the sole beneficiaries of the corporation see: Nicholas Leman (2019) Transaction Man: The Rise of the Deal and the Decline of the American Dream, New York: Farrar, Straus, & Giroux; Nicholas Shaxson (2019) Finance Curse: How Global Finance is Making Us All Poorer, New York: Grove Press.
When researchers substitute reported “costs” for cash flow, they collapse the distinction between accounting categories and economic returns, rendering any subsequent analysis conceptually unsound. This conceptual error becomes even more consequential when applied to the nursing home sector, where cost reports are not designed to measure economic returns and cannot be repurposed to do that. The “cost” construct used in research at issue in this paper collapses operating expenses, related‑party payments, depreciation schedules, and imputed costs into a single accounting artifact that bears no relationship to the cash actually generated by a facility. As a result, research which relies on cost reports to infer financial performance are not measuring returns, adequacy, or shortfall; they are measuring the internal logic of the cost‑reporting system itself. The distinction matters because the gap between reported “costs” and real cash flow is not a subtle discrepancy—it is a structural divergence that becomes unmistakable when examining an actual facility’s financial footprint.
The clearest way to see this structural mismatch is to walk through a real nursing home cost report, where the numbers themselves reveal how far the cost construct diverges from the economic reality of the enterprise. Section II below, is a walkthrough of several 2004 cost report forms submitted by the largest nursing home corporation in the U.S. These forms are revealing: they show how the cost construct obscures rather than illuminates the economic reality of the enterprise.
SECTION II: FORMS G-3, A-1, & A-5: A COST REPORT WALK THROUGH AND ILLUSTRATION WITH A TYPICAL FACILITY
To illustrate how the cost‑report construct obscures rather than clarifies financial performance, we turn to a concrete example: the Big Blue/Riverbend facility owned by the Ensign Group. This facility’s cost report provides a textbook demonstration of the structural gap between reported “costs” and actual economic returns. On paper, the report presents a coherent accounting narrative; in practice, it embeds depreciation schedules, related‑party payments, and imputed capital charges that bear little resemblance to the cash the enterprise generates. By examining the report line by line, we can see how the cost construct inflates capital costs, masks ownership arrangements, and ultimately produces a financial portrait that is analytically incompatible with any measure of return on investment. This walk‑through is not an isolated curiosity—it is a window into the systemic misalignment at the heart of federal cost‑based research.
What makes the Riverbend example especially revealing is that its inflated capital costs do not arise from a single line item but from the interaction of several structural features embedded in the cost‑reporting system. Depreciation is calculated not on the facility’s real asset value but on historical or related‑party book values that can be many times higher than the property’s assessed worth.[i] Lease payments to related real‑estate entities are reported as operating costs even when they function as internal transfers within a vertically integrated ownership chain. Interest expenses may reflect internal loans rather than market‑based financing, allowing owners to generate “costs” by lending money to themselves. Each of these mechanisms is permissible within the reporting framework, yet none of them reflects an actual cash outflow that corresponds to the economic burden implied by the cost report. Together, they create the appearance of a capital‑intensive enterprise where, in reality, the financial structure is engineered to manufacture reportable costs rather than document genuine investment.
[i] The Wyandotte County property database indicates that the building has an assessed valuation or $4 million. A shell company formed as an LLC under Nevada law by the Ensign Group is the listed owner. All of Ensign’s facilities and shell companies are formed as LLCs in Nevada, which, like Delaware is a state with generous tax advantages, high opacity, low liability.
Total costs/expenses reported on line 89 of form A-1 and on the income statement are often misinterpreted as actual expenses, when many non-cash expenses such as depreciation and interest lower taxable income and increase cash distribution to investors. As mentioned earlier, these tax expenditures add value to Medicaid dollars as they flow through the nursing home system.
Every nursing home, every nursing home chain, and every state nursing home system is unique. Nevertheless, the patterns and practices illustrated in the case of Riverbend are common. For instance, line 1, column 5 – capital costs for land and buildings – appears to be highly inflated given the assessed value of the building.
This is an OPCO/PROPCO structured facility. The land and building are owned by an Ensign Group shell company. It is also an OPCO/MANCO structure. The licensee is a shell company through which home office allocations flow. In addition, the facility includes an OPCO/RELPCO component with many of its goods and services purchased from Ensign Group subsidiaries. All of the company’s 400 facilities are LLCs formed in the state of Nevada, which lowers its tax burden and facilitates secrecy.
Table 1 is the 2024 income statement (Form G-3) submitted by the Ensign Group to the Kansas Department of Aging & Disability Services and finalized by the Kansas state auditors. The statement notes net patient revenue (Medicare, Medicaid, and self-pay) of $14,356,777.
Expenses in Table 1 are shown as $13,927,639 – less than net patient revenue even though these expenses include noncash expenses such as depreciation and interest plus payments to related parties, and the home office. So, a large portion of the noted expenses are tax write downs and transfers to related parties at a questionable cost that enhance cash flow. Even with costs that are non-costs folded into expenses shown on the income statement, the facility had a net operating income of 3%.
Table 1: Ensign Group 2024 Form G-3 Income Statement
Tables 2 and 3 are partial views of capital costs (costs for land and buildings) submitted on Form A-1. Line 1, column 5 in Table 3 is the reconciliation with company ledgers for capital costs (land and buildings). It is puzzling that the company’s submitted costs were increased by state auditors by $246,123 as allowable costs (Line 1, column 6). The building is assessed at slightly more than $4 million. Even 25% of a building’s assessed value submitted as capital costs in one year is excessive, but the state increased it to 30%.
A state auditor contacted for an explanation of this extraordinary expense was also perplexed by the state’s action. However, the auditor could not identify a person in the agency who could explain it. This treatment of capital costs for real estate is not uncommon. There is no way of knowing if the costs are due to a triple net lease, plus an interest only loan and accelerated depreciation.
The ASPE report failed to recognize the highly questionable capital costs, home office allocations, and related party payments in their dependent variable construct. Reviews of thousands of facility cost reports suggest that overpriced services and goods are common but ignored by policy makers and researchers.
Table 2: Ensign Group 2024 Form A-1, Lines 1-3, Columns 1-3
Table 3: Ensign Group 2024 Form A-1, Lines 1-3, Columns 4-7
A considerable amount of Medicaid funding flows out of privatized facilities into parent companies through subsidiaries providing goods and services. These subsidiaries are generally limited liability companies. Many are shell companies with no buildings and no employees. Through transfer pricing – pricing higher than fair market value – Medicaid funds are leveraged into increased cash flow.
Tables 4, 5, and 6 show data from the Ensign Group’s submittal of Form A-5, which lists payments to related parties. The company claimed $1,383,636 in payments to their home office and variety of subsidiaries. The Kansas auditor increased that amount by $634,832, which consequently increased allowable costs from related parties to $2,018,568. The public has no way to know the rationale for these decisions by auditors.
The opacity of state auditing and the inaccessibility of government employees making decisions about allowable costs violates the right of the American people to know how their taxes are managed. Secrecy and lack of transparency create conditions for corruption and industry propaganda. The taxpaying public should be able to discern if providers are extracting excessive cash at the expense of patients’ care.
Without transparency, objective, scientific studies cannot be undertaken. Testable hypotheses are only possible in environments where verifiable data is available. The pricing of products and services by subsidiaries of parent/holding companies to its licensed LTC/SKN facility(s) should be compatible with markets. But knowing if that is the case is practically impossible.
Governmental entities and legislators could be considered the main culprits in the opaque, confusing, and in many cases, corrupt practices common in the U.S. nursing home industry. In contemporary business management philosophy, managers are driven by shareholder-return as their guiding ethic. It is not surprising that companies will ignore stakeholder rights and ethical business practices. Hence, the responsibility for honest, ethical management practices rests with regulatory agencies and legislators. They are failing badly.
Table 4: Ensign Group 2024 Form A-5 (Related Parties)
Table 5: Ensign Group 2024 Form A-5 (Related Parties, Columns 2 & 3)
Table 6 Ensign Group 2024 Form A-5 (Related Parties, columns 4–6)
Expenses in the calculation of net operating income include non-cash expenses such as depreciation and interest. Furthermore, transfer pricing by related parties is frequently inflated above market prices. For instance, financiers understand that food purchased in bulk generates mark downs for parent/holding companies (i.e., home offices) but can be priced higher for sale to facility subsidiaries. The same financial machination applies to labor, housekeeping, maintenance, insurance, and other ancillary services.
From a cash flow perspective, an increase in cash flow, and consequently in economic reward for investors, is embedded in costs – otherwise known as expenses. An increase in some highly significant expenses paradoxically increases return on investment, cash and cash equivalents on the balance sheet, and consequently enhances liquidity.
In nursing home finance, cash flow refers to adding value to Medicaid, Medicare, and self-pay funds through financial engineering. In order to enhance cash flow, investors leverage other forms of tax expenditures. For instance, federal and state tax codes allow for write downs for depreciation and interest.
By engaging in tax and lending arbitrage, taxable income can be greatly reduced, thereby increasing availability of cash for distribution to investors, financing acquisitions and other investments, maintenance of liquidity, stock buybacks, and so forth.
Suffice it to say that Medicaid dollars that flow into the privatized nursing home system increase in value through financial engineering not apparent at the facility level. It is practically impossible to measure capital, nay cash, that flows through an opaque network of shell companies with hidden ownership.
SECTION III: TECHNICAL STATISTICAL & METHODOLOGICAL ISSUES
Nursing home care involves matters of life and death. Resources required to insure optimal care often depend on political decisions, which in turn depend upon a technically persuasive argument by advocates. Regression models are often applied to data flowing from complex, dynamic social systems with many interacting subsystems. The output produced by regression algorithms is typically intimidating to professionals untrained in mathematical/statistical fundamentals of intermediate or advanced statistics.
Consequently, the results of even extremely poor modeling designs can silence critique. Parties with a vested interest in an outcome can bully adversaries and undermine open and fair discourse. It is the duty of professionals trained and experienced in methods and statistics to challenge unsuitable research designs harmful to the health and well-being of patients. But too often invalid and misleading modeling goes unchallenged.
ASPE Report Statistical Modeling
The authors presented a multiple ordinary least squares regression model. This report will discuss the following inherent and often interrelated errors and fallacies in the model as specified:
The Simple & Multiple OLS Regression Model Explained
The regression models presented by the authors are most certainly intimidating for lay people without specialized training in intermediate or advanced statistics. Nevertheless, they simply state that variables such as Medicaid payment rates, Medicaid costs, and Medicaid payment-to-cost ratios can be predicted by computer algorithms (mathematical formulas) that calculate the effects of a group of predictor variables. For instance, the variables “ownership type,” “Number of beds,” “Occupancy rate,” “Medicaid payer mix,” “HPRD,” and “star rating” were each weighted based on their effects on the outcome variables.
The ratio between the modeled payment rate to modeled cost was entered as a dependent or outcome variable in a third model. These models are highly questionable for a variety of errors in design. The modeling of costs through specifying “costs” as the outcome/predictor variable when the stated costs on Form A-1, line 89, column 7, as already discussed, are nominal costs but not actual costs. To reiterate, these metrics hide cost reductions that add value to Medicaid.
For professionals with an interest in regression analysis, the multiple regression models discussed are presented below symbolically and with actual results showing the problematic nature of categorized variables such as HPRD, bed size, etc.
The purpose of regression analysis is to explain variance. All cases in an analysis can be expected to vary from the average. When other variables are taken into account (controlled for), different levels of the variable of interest reflect higher or lower averages. The average blood pressure of all U.S. residents may be 120 over 85. However, when controlling for age, the average blood pressure will increase linearly by age. Variance decreases at each level of age. Controlling additional variables will decrease variance even more.
Theoretically, adding additional variables will increase the amount of variance explained and allow for prediction of outcome variables such as “reimbursement rate.” However, specifying the correct variables and all of the interaction between variables in a complex, dynamic social system is not only daunting, but also practically impossible. The fallacy that results is known as reductionism. Reducing a social system as complex as the U.S. nursing home industry to a few variables is illogical.
Regression analysis can baffle lay people and add a patina of science to what is essentially an unscientific study. The following is a technical explanation of modeling with ordinary least squares regression modeling for individuals who would like to work through it. The error rate represented by in the models below is an indication of how well a model predicts the outcome variable, e.g., reimbursement rate or cost. That is a critical diagnostic in all appropriately reported regression models but was ignored in the ASPE report.
Adequately presented regression models include an analysis of error. The distance of an actual value from its predicted value is known as a “residual.” Residuals should be analyzed and graphed. An in-depth discussion of regression diagnostics is beyond the scope of this paper. The point is that a paper with a model of grave implications for the nursing home industry and patients should reflect serious scientific and mathematical standards. The ASPE paper reveals a cavalier attitude toward the power and influence of mathematical models.
A Formal Presentation of Regression Analysis
The Model Specified by the Authors: Problematic Categorization of ratio level measure
ASPE researchers for some unexplained reason categorized equal interval/ratio level measures. They presented no evidence that such a transformation is necessary. In fact, variables such HPRD, bed size are normally distributed and in accord with regression assumptions. For instance, Table 7 and the histogram in Figure 2, indicate that HPRD is normally distributed.
By categorizing continuous variables suitable for entry into a regression model, the researchers increased error in their model and undermined the reliability and validity of predicted rates. The following models are application of the ASPE regression coefficients to two different facilities with slightly different demographics but a significant difference in what the model predicts.
One analysis models a for profit facility with 61 beds, 70 percent occupancy, an HPRD of 3.5, 70 percent Medicaid, and 2 stars. The other models a for profit facility with 60 beds, and 3.4 HPRD. All other variables are the same for each facility.
Although the two facilities are, for all practical purposes, exactly the same, the predicted rates of 186.24 and 183.69 are disturbingly different.
Model I: For profit with 61 Beds, 70 percent occupancy, 70 percent Medicaid, 3.5 HPRD, & 2 stars
Model 2: Same as Model 1 except with 60 Beds & 3.4 HPRD
Categorization of continuous variable is puzzling given the suitability of HPRD and the other equal interval, ratio level measurements for regression analysis. Table 7 and the histogram in Figure 1.
Table 7: 2024 HPRD Descriptive Statistics
Context Variables & Hierarchical Linear Modeling: The Fallacy of Ignoring Nesting
Ignoring “nesting” is a major fallacy in regression modeling. For instance, if students reading scores across a school district were regressed on family income without accounting for each child’s teacher, and school. Students, as the unit of analysis are nested in teacher/classroom, which is in turn nested in school. Without accounting for the effects of the teacher and the school, the outcome variable – reading score – rests solely on the students’ family income. If reading scores were regressed on family income in one classroom or one school, the intercept and slope in might randomly be the same as in another classroom or school, but the probability of that is quite low.
Nursing homes are nested within states. Every state’s nursing home policy, culture, and demographics is unique. For instance, Kansas has 275 facilities with an average of 60 beds. California’s 1200 facilities average 110 beds. The intercept and regression line would be unique for each state. Therefore, a model that accounts for the variance in state intercepts and trend lines is necessary.[
The two-level model – hierarchical linear model or HLM – would look like this:
Two-Level HLM (Random Intercept) Model
Level-1 (Facility Level)
This models how facilities differ within each state
ASPE Cost Adjustments & The Cost Construct
ASPE designed the critically important dependent variable by making four adjustments to the amount of “Net Expense for Allocation” noted on Form A-1 (line 89, column 7). In the first instance, line 89, column 7 is conceptually a poor measure of costs. This particular notation on the form A-1 is not a measure of “cost of caring for Medicaid residents.” It is a post-reclassification, post-adjustment number that includes:
The adjustment for therapy is the least conceptually pure measure of “cost” in the analysis. Indeed, on pages 7-8 and in the footnotes, ASPE acknowledges that line 89 includes categories that do not reflect Medicaid cost of care. They adjust costs to account for the fact that Medicare pays more for therapy than Medicaid. This is an ad hoc adjustment, not a cost-based one.
For instance, therapy costs are notoriously inflated due to upcoding. In 2019, some facilities were transitioning to the patient driven payment model (PDPM) to prevent the ease with which facilities could upcode under the resource utilization groups (RUGs) method of weighting rates by RUGs category. While some states were still moving to PDPM in 2019, many were still measuring case mix through the older RUGs methodology. There is no scientific evidence that PDPM is solving the over-billing problem.
Portioning out therapy hours merely by proportion Medicaid versus Medicare without actual therapy hours accurately assigned to one or the other payor ignores the case mix index for both groups. Medicaid CMI as well as Medicare CMI will vary from facility to facility.
They attempt to remove capital costs (depreciation, interest, lease/rent). But capital costs folded into line 89 are: (1) reclassified, (2) mixed with related-party markups – often inflated, and (3) not separable without facility-level detail.
The ASPE researchers attempted to subtract ancillary service costs (lab, radiology, pharmacy, etc.). But ancillary costs are often transfer payments to related parties. There is no evidence to be found in cost reports of the alignment of transfer pricing with market prices. Sometimes these expenses are embedded in overhead. Sometimes they are reclassified into nursing or administrative lines.
They attempted to exclude costs for services not related to nursing home care (e.g., home health, hospice, outpatient therapy). These costs are rare and insignificant in total costs on line 89.
Can costs reported on facility level cost reports be disentangled for the purposes of measuring the value of Medicaid dollars in cash flow? Perhaps they could be if researchers had access to company ledgers and/or owners were transparent regarding property ownership, internal rates of return and cap rates, loans, and so forth. The same kind of transparency would be necessary in regard to the eclectic mix of related party subsidiaries doing business in insurance, therapy, dietary services, housekeeping, and labor contracting.
SUMMARY
Poorly conceptualized and implemented studies of nursing home finance lead to mismanagement of public funds and often justifies subpar care for patients. Indeed, the ASPE report conclusions are erroneous and biased in favor of for-profit investors. Findings from that unscientific study reinforce a false industry narrative of financial hardship imposed on providers through inadequate Medicaid funding.
Most for-profit nursing home companies are closely held. Their consolidated financial statements are not publicly available. So, industry-wide data are not accessible for verification of cash flow and liquidity. Therefore, scientific claims based on facility-level data cannot be verified. Cash flows through a secretive system of shell companies onto the hidden legers of parent/holding companies remain closeted on financial statements closed to the public.
Evidence from the study of a large publicly listed nursing home corporation with a high percentage of Medicaid patients suggests that the business is quite lucrative for investors.[1] Furthermore, it can be deduced from the review of facility cost reports that real estate assets and transfer payments to related parties and home offices support robust cash flow through parent/holding companies to shareholders. This evidence and deductive logic was ignored by the ASPE researchers.
It is remarkable that full costs on line 89, column 7 of Form A-1 was accepted by the researchers at face value and served as the denominator in the study’s critical Medicaid to cost ratio. A post hoc adjustment was made for therapy services, but there is no means by which that can be verified without patient level data and allowance for upcoding. A guesstimate is not a scientific fact.
The invalid, poorly designed, and inappropriately reported regression models will only serve as a barrier to a full and fair communication process. Regression modeling even at the sophomoric level of this study is intimidating to the uninitiated. Coefficients and statistical significance are well understood by professionals who are trained and experienced in the field of statistics and methods.
Common sense should also tell one that a system as complex as the U.S. nursing home system cannot be reduced to a small number of coefficients. However, if data taken from nursing home cost reports are misunderstood and if Medicaid funds are considered to have a rigid and unchanging value as the flow through shell companies and other entities, the results of regression models will be biased in favor of the industry.
[1] D. Kingsley & C. Harrington (2022), “Financial & Quality Metrics of a Large Publicly Traded Nursing Home Chain in the Age of Covid-19.” International Journal of Health Services,1-13.
[1] https://aspe.hhs.gov/reports/assessing-medicaid-payments-costs-nursing-homes [1] An extensive discussion of cash flow is beyond the scope of this paper. Nevertheless, it must be noted that the income statement of the cost reports (Form G-3) does not inform public policy as it pertains to fair economic returns for profit seeking providers. Net operating income – net patient revenue minus expenses – is often misinterpreted as profit or the ultimate return on investment. [1] In business management theory as it has evolved since the 1970s, managers are “agents” of shareholders and have no ethical responsibility for stakeholders such as customers, patients, citizens, and communities. For two excellent discussion of the transition from a 1950s-60s philosophy of corporate public responsibility to today’s ethic of shareholders as the sole beneficiaries of the corporation see: Nicholas Leman (2019) Transaction Man: The Rise of the Deal and the Decline of the American Dream, New York: Farrar, Straus, & Giroux; Nicholas Shaxson (2019) Finance Curse: How Global Finance is Making Us All Poorer, New York: Grove Press. [1] The Wyandotte County property database indicates that the building has an assessed valuation or $4 million. A shell company formed as an LLC under Nevada law by the Ensign Group is the listed owner. All of Ensign’s facilities and shell companies are formed as LLCs in Nevada, which, like Delaware is a state with generous tax advantages, high opacity, low liability. [1] See: Robert Bickel (2007) Multilevel Analysis for Applied Research. New York: The Guilford Press; Jos W.R. Twisk (2006) Applied Multilevel Analysis. New York: Cambridge University Press; [1] D. Kingsley & C. Harrington (2022), “Financial & Quality Metrics of a Large Publicly Traded Nursing Home Chain in the Age of Covid-19.” International Journal of Health Services,1-13.Category: Uncategorized