Challenging a Fundamental Proposition of Patient-Centeredness

Abstract

This investigation challenged the proposition that physician patient-centeredness influences patients’ experience-of-care (PEC). A theory-driven, three-factor, multigroup structural equation modeling design, using asymptotic-distribution-free and bootstrap estimation, with two national random and 5,000 bootstrap samples challenged the proposition’s plausibility, measurement invariance, replicability, robustness against a competing model, and coherence with theory. The model fit [χ²(39) = 28, p =.900, RMSEA = .001, p = 1.00, CFI = 1.00], explaining 81 percent of PEC’s variance; the proposition was invariant across samples, held against the competing model [χ²Δ(7) = 7.82, p = .97]; cross-validated against estimates from the 5,000 bootstrap samples; and agreed with theory. One standardized increase in patient-centeredness increased PEC, likelihood of recommending, and care ratings by .807, .765, and .771. Results converged in sustaining the plausibility of the proposition.

INTRODUCTION

While patient-centeredness and patient-centered care are related, they are not the same. Patient-centeredness is the underlying ability of healthcare providers that influences the quality of their communication and interaction with patients, and consequently the patient’s resulting experience-of-care. As an ability, it is the formative precursor of a provider’s method of diagnosis, treatment and care. As such, patient-centeredness complements technical clinical competence, and it positively influences and adds value to patients’ experience-of-care. It is the important component of healthcare that responds to patients’ needs, preferences, and values (Cooper, Holmes, Seid, Gonzalez, & MacDonald, 2019; Medicine, 2001a). Alternatively, a healthcare providers’ patient-centered care refers to care behavior(s) that result from their patient-centeredness ability.

As a rule, patient-centeredness and technical clinical competence are necessary and sufficient conditions for a patient’s positive experience-of-care, as well as other outcomes, like trust, compliance, loyalty, engagement, likelihood to recommend, ratings of care, quality, acceptance of diagnoses and treatment plans, motivation, reduced litigation etcetera (S. Aragon et al., 2012; S. J. Aragon, McGuinn, Bavin, & Gesell, 2010; S. J. Aragon, Richardson, Lawrence, & Gesell, 2013; S.J. Aragon, Sherrod, McGuinn, & Gesell, 2019; Chou & Cooley, 2020).

The purpose of this investigation is to empirically challenge the proposition that physician patient-centeredness influences patients’ experience-of-care. To challenge this proposition, patient-centeredness and patients’ experience-of-care were modeled as independent exogenous and dependent endogenous constructs respectively, and then vigorously tested across two national random samples of emergency department patients. For cross-validation, the proposition was also tested across 5,000 bootstrap samples, randomly selected from the original national random samples with replacement.

Background

Given increasing health disparities, lack of equity, unsafe practices, medical errors, and untimely care, the Institute of Medicine (IOM, 2000; 2001a; 2001b; 2003) and others (Cooper, Holmes, Seid, Gonzalez, & MacDonald, 2019; Institute of Medicine, 2000, 2001a, 2001b, 2003) recommended that providers adopt a more patient-centered, equitable, effective, timely, and safe model of care, where decisions were informed by patients’ needs, preferences, and values. Nationally, healthcare systems, their hospitals, and management teams responded by implementing a range of new patient-centered care programs (Center, 2020; Clinic, 2020a, 2020b; Health, 2020; Institute, 2020a, 2020b).

However, despite these new programs, the achievement of a patient-centered healthcare system has a long way to go. While some improvement has occurred, serious problems remain (Quality, 2019). Avoidable harm continues to cause patient morbidity and mortality (Makary & Daniel, 2016; Quality, 2019; Sargent & Waldman, 2019). The uninsured and poor, particularly African Americans, American Indians and Alaskan Natives, and Native Hawaiians/Pacific Islanders, continue to experience disparities in the delivery of care and outcomes. The healthcare system’s safety, effectiveness, access, and cost all need transformational improvement (AHRQ, 2018).

Federal Response

In response, the Centers for Medicare & Medicaid Services (CMS) instituted a national policy imperative called the Value-Based Purchasing Program (VBP) to create a more patient-centered healthcare system, prevent and overcome disparities, and improve clinical outcomes, experience-of-care, safety, efficiency and costs for Medicare inpatients. VBP is pivotal because it pays hospitals, and therefore their managements to improve physicians’ and nurses’ communication, courtesy, respect, listening, explanations, and responsiveness, and it penalizes them if patients’ experience-of-care, clinical outcomes, safety, efficiency and costs scores do not meet pre-established performance benchmarks and thresholds (Carter John & Silverman Fred, 2016; CMS, 2020; Guadagnino, 2012; Mazurenko, Collum, Ferdinand, & Menachemi, 2017; Stacy, 2016; Thiels et al., 2016).

The VBP paradigm’s financial incentives and penalties have the effect of monetizing patient-centeredness and patients’ experience-of-care scores, encouraging hospital managers to improve patients’ experience, clinical outcomes, safety, efficiency and cost (CMS, 2020; Sargent & Waldman, 2019; Simsir & Altindis, 2019; VanLare, Moody-Williams, & Conway, 2012). CMS also expects VBP incentives to increase evidence-based care and transparency for patients (CMS, 2020; Lasater, Germack, Small, & McHugh, 2019; Lee, Venkataraman, Heim, Roth, & Chilingerian, 2020; Makary & Daniel, 2016; Qi, Luke, Crecelius, & Maddox, 2019; Simsir & Altindis, 2019; Ternay & Mgma, 2019).

The use of value-based care incentives and penalties is gaining momentum. CMS advised its state Medicaid directors to adopt and implement patient-centered, value-based policies to address disparities and the social determinants of health (CMS, 2020).

Counterfactuals of Patient-Centeredness

Despite the positive efforts to transform the healthcare system into one that is more patient-centered, there have been a few naysayers. Some contend that the quality of physician-patient communication is a function of the patient’s education and language competence (Aelbrecht et al., 2019; Valentijn et al., 2018). Others challenge the benefit of shared decision-making, a corner-stone of patient-centered care, proposing for example, that family family values and preferences be excluded from decision-making in pediatric intensive care units (Richards et al., 2018). Likewise, doubt has been cast on the importance of physician personality and behavior on patients’ experience-of-care (Boerebach et al., 2014). Others have questioned the value of transparency and giving patients access to their medical records (Davis, Menon, Parrish, Sittig, & Singh, 2014). The obligations of patient-centered care have also been questioned when they are inconvenient or challenging (Fiester, 2012).

Skepticism should be considered in balance. Skepticism is a valuable component of critical thinking, even when it runs counter to an established body of generally accepted experience and evidence. The Greek word skepsis suggests investigation by to someone who is unconvinced and seeking truth. As such, skepticism will always play an important role in advancing knowledge. Consistent with the need for both investigation and balance in the face of skepticism, the following describes the investigation’s methods to determine the validity and reliability of the proposition that physician patient-centeredness influences patients’ experience-of-care in the emergency department.

METHODS

Purpose and Model

The purpose of this investigation was to challenge the proposition that physician patient-centeredness influences patients’ experience-of-care. After modeling the proposition and estimating its directed relationships (population parameters), the results were subjected to evidence-based tests with sufficient power to sustain or falsify the proposition’s validity and reliability.

Figure 1 represents the modeled proposition that physician patient-centeredness influences patients’ experience-of-care in emergency departments, with waiting time as a mediator.

Figure 1:

Hypothesized Model Specifying the Proposition that Physician Patient-Centeredness, Mediated by Waiting, Influences Emergency-Room Patients’ Experience-of-Care

Model Measures

Table 1 presents the model’s measures: Physician Patient-Centeredness (PPC), Patient-Centeredness of Waiting (PCW), and Patients’ Experience-of-care (PEC) are its key constructs, where the dependent observed variables used to measure them (c2, c75, c5, c1, F1, F68, and F4) are informed by patients’ ratings on a 5-point scale, from very poor (1) to very good (5). For example, PPC measures the patient-centeredness ability of ER physicians, as a linear combination of their courtesy, listening, and informativeness as evaluated by their patients.

Table 1:

Measures and Descriptions

PPC	Physician Patient-Centeredness
c2	Physician courtesy
c75	Physician listening
c5	Physician kept patient informed
PCW	Patient-Centeredness of Waiting
c1	Waiting time before seeing physician
F1	Informed about delays
PEC	Patients’ Experience-of-care
F68	Overall rating of care
F4	Likelihood of recommending the emergency department

Theory

The Primary Provider Theory (Aragon, 2000) guided the design of this investigation (Fig. 2). It holds that providers’ patient-centeredness influences patients’ experience-of-care. The theory is generalizable and can accommodate a full range of research hypotheses, patient populations, healthcare settings, and providers, including physicians, nurses, allied health practitioners, dentists, phlebotomists, podiatrists, and physical therapists, and others.

Figure 2.

Primary Provider Theory’s Nomological Network of Patient-Centeredness Constructs (Large Circles) and their Effects (Arrows) on each other and Patients’ Experience-of-Care the Dependent Construct. Dis 1 and 2 Represent all Omitted Causes and Random Measurement Error. Note. Not represented are the observed variables that operationalize each construct.

The theory is grounded in guiding principles that favor and support all patients and their experience-of-care. For example, one principle holds that clinical competence is a necessary but insufficient condition of patients’ experience-of-care, maintaining that technical-clinical-competence alone is insufficient because the actual delivery of care also requires communication and interaction and with patients. Another principle holds that healthcare providers are uniquely responsible for the quality of their communication and interaction with patients. Another maintains that, above all else, healthcare providers are ethically responsible for protecting the best interests of their patients.

The theory holds that patient-centeredness is a formative ability of healthcare providers that affects the quality of their interaction and communication with patients. As such, the patient-centeredness of the primary provider has direct and indirect effects on patients’ experience-of-care. Lastly, the theory holds that patients are the best judges the patient-centeredness of their health providers and their eventual experience-of-care.

The theory was chosen because of its generalizability, robustness, and relevance to the investigation’s purpose and applicability to patient-centered hypotheses, outcomes, various healthcare settings, and practitioners (Aragon, 2000; Aragon & Gesell, 2003; Aragon et al., 2010; Aragon et al., 2012). Appendix B presents a graphical representation of the theory’s interrelated network, key guiding principles, and implied structural interrelationships.

Design

To challenge the hypothesized model (Fig. 1), the investigation applied a theory-driven, three-factor, multigroup, structural equation modeling design, with parametric asymptotic distribution-free (ADF) and nonparametric bootstrap estimation, two cross-sectional national random samples of emergency patients (N_t = 284; N_c = 294), and 5,000 bootstrap samples. The model’s measurement reliability and validity were confirmed before its incorporation into the full structural model (Bollen, 1989). Evidence-based, within and between-group tests were employed to establish the model’s goodness of fit, replicability, and measurement invariance across all samples. The design also challenged the hypothesized model’s robustness against a competing model, and the means, standard errors, and confidence intervals generated from 5,000 bootstrap samples. Also assessed were the model’s ability to explain patients’ experience-of-care, and its coherence with theory. Lastly Hochberg’s method was used to control for multiplicity or familywise error and to reduce the risk of finding parameters significant, if they were not (Cribbie, 2007; Hochberg, 1988). Mardia’s measure of multivariate kurtosis was used to test the data for distributional violations and to inform the estimation method (Arbuckle, 1994; Mardia, 1970; Tsamardinos, Greasidou, & Borboudakis, 2018). The model was specified with sufficient degrees of freedom (df = 39), precision, and power (α = .90) to reject or sustain its plausibility (MacCallum, Browne, & Sugawara, 1996). The investigation was conducted pursuant to Winston-Salem State University’s Institutional Review Board (IRB# 2986–07-0035).

Data

The study included 578 unique, randomly selected, full-information patients from Press Ganey’s national emergency department database. A proxy of the national population, this database contains over 1,477,835 patients who visited 1,383 emergency departments across 49 states. Their responses were collected with Press Ganey’s Emergency Department Survey, which is listed by the National Quality Measures Clearinghouse and has a reliability of a = .92 (Hall & Press, 1996). The survey includes patients’ ratings of emergency department physicians (which are indirect measures of physician patient-centeredness) and patients’ related experience-of-care on a 5-point scale from very poor (1) to very good (5). Patients were randomly assigned to test (N_t = 284) and control (N_c = 294) samples. Patients’ age ranges (M = 50.2, SD = 23.56) and time spent in the emergency department (M = 5.33, SD = 5.71) were similar across samples. Press Ganey de-identified all data before providing it to the researchers. Appendix C provides the moments for these samples.

Model Specification

In accordance with the hypothesis that physician patient-centeredness influences patients’ experience-of-care both directly and indirectly through waiting (Fig. 1; Table 1), the model was specified with 29 parameters (i.e., regression weights or effects and variances) and 39 degrees of freedom. To test the model’s measurement invariance and replicability, 17 parameters were fixed to be equal, across the national random test and control and 5,000 bootstrap samples. Appendix D provides additional specifications for measuring the model’s structural linear relationships and implied causal equations.

Model Estimation

Mardia’s measure was used to assess the multivariate normality of the data. Both samples proved to be multivariate kurtotic (N_t = 284, kd = 101, p < .05 and N_c = 294, kd = 89, p < .05). Given this nonnormal character of the data, parametric asymptotic distribution-free estimation (ADF) was used to estimate all model parameters. These parameter estimates were cross-validated with the nonparametric estimates (means, standard errors, and confidence intervals) generated by 5,000 bootstrap samples to discern any parameter-estimation bias.

Model Evaluation

The model’s plausibility was determined by the convergence of empirical evidence from tests of (a) goodness of fit; (b) parameter invariance and replication across samples; (c) robustness against a competing model; (d) cross-validation with the means, standard errors, and confidence intervals of 5,000 bootstrap samples; (e) the magnitude of the variance of patients’ experience-of-care explained; and (f) coherence with the Primary Provider Theory in terms of the direction, size, and significance of parameter estimates.

RESULTS

Measurement Model

Reliability and Construct Validity (convergent and discriminant).

Table 2 shows evidence supporting the model’s construct reliability (CR) and average variance extracted (AVE), which exceed the standard threshold of .70 (Hair, Anderson, Tatham, & Black, 1998). AVE and average squared interconstruct correlation (ASC) values greater than one also support the model’s discriminant validity. Taken together, these measures support the measurement model’s construct validity and appropriateness for incorporation into the full structural equation model for estimation.

Table 2:

Construct Composite Reliability (CR), Average Variance Extracted (AVE), and Average Squared Interconstruct Correlations

Constructs	CR	AVE	ASC	AVE/ASC
Patients’ Experience-of-care	0.955	0.906	.727	1.25
Patient-Centeredness of Physician	0.963	0.898	.651	1.38
Patient-Centeredness of Waiting	0.848	0.736	.730	1.01

CR = construct composite reliability, AVE = average variance extracted, ASC=average squared interconstruct correlations.

Structural Model

Goodness of fit.

The model fit well across samples (χ²(39) = 28.08, p = .90, RMSEA = .001, p = 1.00, RMSEA 90% CI = .00 - .01, CFI = 1.00). Converging with this result, were the model’s standardized residual covariances. In sufficiently large samples, for correct models, standardized residual covariances have a standard normal distribution, and therefore should be less than 2.0 in absolute value. All standardized residual covariances were less than 2.0, equating to a standardized root mean residual of SRMR = .032, which also reflects a correct model (Hu & Bentler, 1999).

The resulting total standardized effects of physician patient-centeredness on emergency patients’ experience-of-care, likelihood of recommending, and ratings of care were .807, .765, and .771 respectively, while the analogous effects of waiting time were .652, .618, and .623 (see Table 4). Together, physician and waiting patient-centeredness accounted for 81 percent of patients’ experience-of-care (R² =.81, 90% CI [.724, .851], p = .002).

Table 4:

Total Effects (Standardized)

	Physician Patient-Centeredness	Patient-Centeredness of Waiting
Patient-centeredness of waiting	.794	.000
Patients’ experience-of-care	.807	.652
Likelihood of recommending	.765	.618
Ratings of care	.771	.623

Measurement invariance and replication across samples.

Results supported the model’s invariance and the replication of its effects across samples (χ²Δ(7) = 7.82, p = .97). Total effects also held across samples. These results support the proposition that physician patient-centeredness positively influences patients’ experience-of-care, mediated by waiting time, in emergency departments. They also support the generalizability of the challenged proposition across emergency departments and the likelihood that one variance-covariance matrix can represent the national population of emergency department patients, where experience-of-care is influenced by physician patient-centeredness mediated by waiting time.

Cross-validation with means, standard errors, and confidence intervals from 5,000 bootstrap samples.

First derived parametrically using ADF estimation methods, the model’s effects were re-estimated post hoc to challenge their accuracy and precision against their nonparametric bootstrapped counterparts. In addition to testing the model’s ability to cross-validate against these nonparametric means, standard errors, and confidence intervals, the model’s parameter-estimate bias was assessed as the difference between the nonparametric estimates from bootstrap estimates and parametric ADF estimates.

Table 5 shows that the model’s ADF estimates (β₂) cross-validated well with the bootstrap estimates (β₃), revealing negligible parameter-estimation bias (β² - β³). For example, β₃, the bootstrap effect of “provider patient-centeredness” on patient’s ratings of “physician courtesy” (PPC → C2) is .948, while the model’s ADF estimate β₂ is .941. The difference β² - β³ reflects the amount of estimation bias (.948 – .941 = .007).

Table 5:

Standardized Structural Estimates, Bootstrap Means, Standard Errors, Parameter Bias, Confidence Intervals, Probabilities, and Squared Multiple Correlations

Parameter	β2	β3	SE	Bias	90% CI	p	R2
PPC → PCW	.794	.801	.035	.006	[.724, .841]	.002	.631
PPC → PEC	.289	.292	.084	.003	[.132, .407]	.009	.808
PCW → PEC	.652	.658	.078	.006	[.533, .782]	.001	.652
PPC → C2	.941	.948	.012	.007	[.906, .953]	.026	.885
PPC → C75	.958	.961	.011	.003	[.933, .971]	.004	.918
PPC → C5	.944	.945	.013	.001	[.917, .963]	.001	.891
PEC → F68	.955	.965	.012	.010	[.922, .966]	.046	.912
PEC → F4	.948	.951	.014	.002	[.920, .968]	.002	.899
PCW → c1	.815	.811	.027	−.003	[.768, .857]	.001	.663
PCW → F1	.899	.906	.020	.007	[.857, .924]	.005	.809

β₂=standardized estimates; β₃= standardized bootstrap estimates across 5,000 samples; SE= standard error of β³; Bias= β² - β³; 90% CI=bootstrap bias corrected 90% confidence intervals; p=two-tailed probability; R²=squared multiple correlations.

Competing model plausibility challenge.

The model’s goodness of fit, with parameters restricted to equality across samples, was compared post hoc to the competing model’s fit, with parameters unrestricted across samples.

Akaike’s information criterion (Akaike, 1973) and Browne-Cudeck’s criterion (Browne & Cudeck, 1989) were used to compare the fit of the competing models. Lower AIC and BCC values indicate the more correct model. Since the hypothesized model was a nested, more restricted version, of the competing model, a chi-square difference test (χ²Δ) informed the contrast [χ²Δ(7) = 7.82, p = .97]. Table 6 shows that the hypothesized model’s fit to be superior, given its lower AIC and BIC values and the aforementioned chi-square difference test.

Table 6:

Comparison of Models’ Chi-square, Chi-square/df, AIC, BCC, and Chi-square Difference Test Results

Model	df	χ2	p	χ2/df	AIC	BCC
Competing model (effects specified unequal across samples)	22	.776	.555	.930	88.46	90.40
Hypothesized model (effects specified equal across samples)	39	28.28	.898	.725	62.28	63.25
Chi-square Difference Test	7	7.045	.900	.460

df = degrees of freedom; χ² = chi-square; χ²/df = chi-square per degree of freedom; AIC = Akaike information criterion; BCC = Browne-Cudeck criterion.

Ability to explain the variance of patients’ experience-of-care.

Accounting for 81 percent of its variance across all samples, the model’s ability to explain patients’ experience-of-care was considerable and significant (R² =.81, 90% CI [.724, .851], p = .002).

Coherence with theory.

Model results corroborated central and sub-postulates of the Primary Provider Theory that provider patient-centeredness directly and indirectly influences patients’ experience-of-care, mediated by waiting. In addition, the model’s effect sizes were consistent with the direction, size, and significance of Primary Provider Theory predictions (Figs. 1, 3; Tables 4, 5).

Figure 3:

Standardized Effects from Construct to Construct and from Construct to the Observed Variables (Rectangles), they were Specified to Cause. R²s are Represented on the Edge of each Rectangle. For example, the .88 on rectangle c2 indicates that PPC causes 88% of the variance of c2 (physician courtesy). The β = .94 on the arrow from PPC to c2 reflects the standardized effect of PPC on physician courtesy (c2). The .81 on the PEC construct reflects that PPC and PCW together account for 81% of the variance of PEC. The smallest circles represent measurement error and structural disturbances in the model. Dis 2 accounts for any omitted causes and random measurement error in the model’s prediction of PEC.

Model multiplicity control.

Although the model’s causal relationships, parameters, and significance levels were cross-validated in two samples and with their counterparts from 5,000 bootstrap samples, additional control was taken to reduce the risk of falsely finding parameters significant. Like the bootstrap estimates, the Hochberg estimates in Table 7 confirm the significance (α < .05) of the model’s parameters with negligible inflation (Cribbie, 2007).

Table 7:

Hochberg Multiplicity Control Estimated P-Values

Model Hypothesis	Hochberg P-Values	Model P-Values	P-Value Inflation
PPC → PCW	0.012	0.002	0.010
PPC → PEC	0.027	0.009	0.018
PCW → PEC	0.008	0.001	0.007
PPC → C2	0.046	0.026	0.020
PPC → C75	0.020	0.004	0.016
PPC → C5	0.008	0.001	0.007
PEC → F68	0.046	0.046	0.000
PEC → F4	0.012	0.002	0.010
PCW → c1	0.008	0.001	0.007
PCW → F1	0.020	0.005	0.015

PPC → PEC reflects the hypothesis that physician patient-centeredness (PPC) affects patients’ experience-of-care (PEC). The related p-values represent the significance of the influence or effect of PPC on PEC, where p-value inflation equals Hochberg-estimated minus model-estimated p-values.

DISCUSSION

This investigation was designed to rigorously challenge the fundamental proposition that provider patient-centeredness positively influences patients’ experience-of-care. Evidence converged supporting the proposition and its generalizability across samples of emergency room patients, where physician patient-centeredness significantly influenced the patients’ experience-of-care, their likelihood to recommend, and their ratings of care. The causal structure of the proposition and its effects were invariant and replicated across two national samples of emergency room patients and the means, standard errors, and confidence intervals generated by an additional 5,000 bootstrap samples. The proposition was further sustained when it was compared to a competing proposition of no generalizability. Ultimately, the patient-centeredness of emergency department physicians explained 81 percent of the variance of patient’s experience-of-care, mediated by waiting, and the results matched the causal predictions of the Primary Provider Theory. This evidence supports patient-centeredness as an underlying ability of providers that effects the quality of their communication and interaction with patients. Patient-centeredness is an ability that adds value to technical clinical competence, and it is driver of patients’ experience-of-care.

Who are the providers?

At the local level, healthcare providers are conventionally recognized as physicians, nurses, allied health practitioners, and others who provide direct care to individuals. At the population level however are healthcare corporations, hospital systems, their hospitals and chief executive officers (CEOs) healthcare providers as well? The answer is yes. As physicians are licensed to practice medicine, hospitals are licensed and therefore bound to provide largescale healthcare services to entire populations and communities. These licensed hospitals, run by their CEO’s are required to produce and provide a comprehensive range of healthcare services for the populations they serve according to legally enforceable standards of quality. The range of services include, but are not limited to, health facilities, diagnostic equipment, technology, qualified physicians, nurses, allied healthcare practitioners and technologists, and a myriad medical procedures and tests for entire populations and communities. These services must comply with standards set by their corporations as well as federal, state, and local canons. On a day-to-day basis, powerful CEOs are responsible and accountable for the overall quality of their hospitals to their boards. As such, they have a responsibility and binding employment imperative to maintain and control, directly and indirectly, the service quality, clinical quality, and financial quality of their hospitals, which includes all its employed medical professionals and staff. In addition, their code of ethics requires that CEO’s both identify and meet the healthcare needs of the community they serve (ACHE, 2017). Now, achievement the Value-based Purchasing Program’s patient engagement, clinical, safety, and efficiency standards require CEO’s, their management teams, and the entire hospital staff to be more patient-centered to improve patients’ experience-of-care. While CEOs’ abilities include leadership, communication, professionalism, knowledge of the healthcare environment, and business competence (ACHE, 2020), the best chief executives are also patient-centered, as they expect of their hospitals’ physicians, nurses, allied health practitioners, and staff.

What can happen when a powerful hospital CEO is not patient-centered?

The answer was tragically illuminated by the Phoenix Veterans Administration Health Care System (PVAHCS) case, when it was announced by chair of the House Committee on Veterans’ Affairs that PVAHCS’ CEO and other executives had approved the alteration and destruction of records to hide patient abuse, suggesting that as many as 40 veterans may have died as a result (Inspector-General, 2014b). A firestorm ensued, and in the aftermath, PVAHCS’ CEO as well as the VA’s CEO, the Secretary of the Veterans Administration, resigned and/or were terminated. In response, Congress passed legislation in an effort to create a more patient-centered VA healthcare system. Despite this however, there have been calls for privatization of the VA, allowing veterans to select their own physicians, and hospitals. Patient reports suggesting a systemic lack of patient-centeredness raised a basic question: should the VA provide healthcare at all or serve only as a fiscal intermediary, paying for the care veterans receive elsewhere (Bogardus, 2016; Inspector-General, 2014a, 2014b; Longman & Gordon, 2017; Robbins, 2014).

Limitations and Future Directions. It is noted that data for this investigation were obtained using one survey that could potentially contribute to common method variance. In addition, the manifest variables used to measure physicians’ patient-centeredness ability are not exclusive. There are numerous variables that influence the quality of a physician’s patient-centeredness, that is, their communication and interaction with patients. And, the ratings of patients who responded to the survey could differ from those of non-respondents.

The study offers a paradigm for measuring and improving emergency department patients’ experience-of-care via physician patient-centeredness as well as their courtesy, listening, and information transmission to patients. As a precursor of the quality of their communication and interaction with patients, providers’ have numerous patient-centered behavioral choices. In this light, healthcare systems and their providers must ensure that patients are free and know how to communicate their concerns and preferences (McGuinn, 2021). As healthcare systems expand and acquire increasing size and power, their risk of becoming less patient-centered and indifferent to patients’ needs, preferences, and values increases. Given this peril, all healthcare providers, including hospital boards and their CEO’s must retain their zeal for serving the best interests of their patients with zero tolerance for any abuse of power that may compromise patients, including and discriminatory practices that may lead to disparities.

CONCLUSION

In conclusion, the results advocate for the continued implementation of patient-centeredness reform in emergency departments and the healthcare system. By increasing the number, scope, and influence of patient-centered providers in hospitals, and achieving the Value-Based Purchasing Program’s patients experience-of-care benchmarks, hospital boards, hospital systems, hospitals, and their CEOs and managers, physicians, nurses, and allied health practitioners can achieve a more patient-centered healthcare system for every patient.

Table 3:

Model Goodness-of-Fit Thresholds and Results

Criteria	Thresholds	Results
Chi-square test	χ2 ns	χ2 p = .90
Root mean sq error of approximation	RMSEA = ≤ .05, ns	RMSEA = .001, p = 1.0
Comparative fit index	CFI ≥ .95	CFI = 1.0
Standardized residual covariances	SRC ≤ 2	≤ 2
Parameter estimate bias	bias ≤ .02	≤ .02
Chi-square/degree of freedom	χ2/df < rival	χ2/df < rival
Akaike information criterion	AIC < rival	AIC = 62.28, rival = 88.46
Browne-Cudeck criterion	BCC < rival	BCC = 63.25, rival = 90.40

APPENDICES

Appendix A: Centers for Medicare & Medicaid Services (CMS) Value-Based Purchasing Thresholds and Benchmarks per Performance Domain

Appendix B: The Primary Provider Theory

Developed during his years of service in hospital administration, Dr. Stephen Aragon first published The Primary Provider Theory in 2000. It is an evidence-based, generalizable Theory specifying how the patient-centeredness of healthcare providers influences patients’ experience-of-care and other outcomes (S.J. Aragon, 2000) like trust, compliance, patient loyalty, patient engagement, likelihood to recommend, patients ratings of care and quality, delivering diagnoses and treatment plans, patient motivation, reduced litigation, and more. Based on years of observations, interviews, focus groups, and analyses of patient surveys, the validity and reliability of the theory’s causal relationships (below) have been tested across patient populations, healthcare providers, and settings. The Theory informed the Provider Patient-Centeredness and Disparities Outcome Measurement Initiative, 5-year grant funding empirical research to improve patients’ experience-of-care and to reduce disparities among diverse populations, women, and children (NIH/NCMHD 5-P20-MD002303–03).

Abstract

The Primary Provider Theory maintains that technical clinical competence is a necessary but insufficient condition for achieving positive patient experience-of-care because the delivery requires communication and interaction with patients. This important component of care, called patient-centeredness, is the formative ability of providers that affects the quality of their communication and interaction with patients, and therefore their experience-of-care and other outcomes. The Theory holds that outcomes like patients’ experience-of-care, like trust, compliance, loyalty, engagement, likelihood to recommend, ratings of care, delivering diagnoses and treatment plans, motivation, reduced litigation, and other desirable outcomes, are described by a nomological network of interrelated constructs including the patient-centeredness of the primary provider, and secondarily the patient-centeredness of provider’s associates and waiting time. In the network, these constructs are hierarchically related to patients’ expectations, where the primary provider has the greatest clinical utility and therefore value to patients. Informed by patients’ expectations, in their ratings of providers, the theory can be generalized over a wide range of patient populations, healthcare settings, and providers, including physicians, nurses, allied health practitioners, dentists, phlebotomists, podiatrists, and others.

Principles

1. Above all else, healthcare providers are responsible for protecting the best interests of their patients.

2. Clinical competence is a necessary but insufficient condition of positive patient experience-of-care. Note. Technical clinical competence alone is insufficient because the delivery of care requires communication and interaction with patients.

3. Healthcare providers are solely responsible for the quality of their communication and interaction with patients.

4. Patient-centeredness itself is the formative ability of healthcare providers that affects the quality of their communication and interaction with patients, and therefore outcomes.

5. Together, clinical competence and patient-centeredness are necessary and sufficient conditions of positive patient experience-of-care.

6. Patients are the best judges of the patient-centeredness of their healthcare providers and their experience-of-care.

A generalizable theory of how the patient-centeredness of healthcare providers influences patient experience-of-care and other desired outcomes.

Implied Causal Relationships

$$\text{PEC}=f\left(\text{PPC},\text{PCA},\text{PCW},di{s}_1\right)$$ (1)

$$\text{PCA}=f\left(\text{PPC},di{s}_2\right)$$ (2)

$$\text{PCW}=\text{f}\left(\text{PPC},di{s}_3\right)$$ (3)

Implied Linear Relationships

$$\text{PEC}=(effect)\text{PPC}+(effect)\text{PCA}+(effect)\text{PCW}+(effect)di{s}_1$$ (4)

$$\text{PCA}=(effect)\text{PPC}+(effect)di{s}_2$$ (5)

$$\text{PCW}=\left(effect\right)\text{PPC}+\left(effect\right)di{s}_3$$ (6)

Appendix C

Test Sample Covariances and Means
	F1	c1	F4	F68	C5	C75	C2
F1	1.301
c1	.893	1.143
F4	.863	.733	1.013
F68	.766	.650	.808	.786
C5	.699	.593	.660	.586	.826
C75	.662	.562	.626	.556	.697	.720
C2	.576	.489	.545	.483	.607	.575	.565
Means	3.993	3.972	4.313	4.324	4.335	4.391	4.440
SD	1.177	1.089	1.031	0.951	.997	0.909	0.853
Control Sample Covariances and Means
	F1	c1	F4	F68	C5	C75	C2
F1	1.301
c1	.893	1.143
F4	.863	.733	1.013
F68	.766	.650	.808	.786
C5	.699	.593	.660	.586	.826
C75	.662	.562	.626	.556	.697	.720
C2	.576	.489	.545	.483	.607	.575	.565
Means	4.051	4.129	4.388	4.425	4.425	4.459	4.537
SD	1.18	1.088	1.028	.897	.923	.880	.782

Appendix D: Additional Model Specifications, Hypothesized Model

A subproposition of the Primary Provider Theory specifying that physician patient-centeredness, mediated by waiting, influences emergency-room patients’ experience-of-care.

Measurement Model

Taken together, equations (3) – (9) below specify the measurement model and the linear relationships among its constructs (PPC, PCW, and PEC) and the variables measuring them (c2, c75, c5, c1, F1, F68, and F4).

In equations (3), (4), and (5), the variables c2, c75, and c5 capture patients’ ratings of their physician’s 1) courtesy, 2) listening, and 3) informativeness, which, in the model, are predicted by PPC, where v1, v2, and v3 capture their measurement error.

$$\mathrm{c}2=f\left(\mathrm{PPC},\mathrm{v}1\right)$$ (3)

$$\mathrm{c}75=f(\text{PPC},\text{v}2)$$ (4)

$$\text{c}5=f(\text{PPC},\text{v}3)$$ (5)

Predicted by PCW, c1 and F1 capture patients’ ratings of the patient-centeredness of their 1) waiting time and 2) how well they were informed while waiting, where v5 and v6 capture their associated measurement error.

$$\text{c}1=f(\text{PCW},\text{v}5)$$ (6)

$$\text{F}1=f(\text{PCW},\text{v}6),$$ (7)

PEC predicts F68 and F4, which capture patients’ ratings of their 1) overall care and 2) their likelihood of recommending the service, where v10 and v9 is measurement error.

$$\text{F}68=f(\text{PEC},\text{v}10)$$ (8)

$$\text{F}4=f(\text{PEC},\text{v}9),$$ (9)

Structural Model

Taken together, equations (1) and (2) specify the structural model: the relationships among PPC, PCW, and PEC, and the patient ratings they predict (c2, c75, c5, c1, F1, F68, and F4). Equation (1) specifies that together PPC and PCW cause and predict PEC. Equation (2) specifies that PPC influences and predicts PCW. The variables dis1 and dis2, called model disturbances, capture variance related to omitted causes and random measurement error.

$$\text{PEC}=f(\text{PPC},\text{PCW},\text{dis}2)$$ (1)

$$\text{PCW}=f(\text{PPC},\text{dis}1)$$ (2)