The effect of patient quality measurements and HCAHPS patient satisfaction on hospital reimbursements

2.1Hospital value-based purchasing programs (VBP)

Beginning in 2003, CMS developed a process to provide financial incentives to hospitals which produce higher quality, more cost-effective care. The most prominent example of this process is a pilot project in 2003 based on pay-for-performance. The Premiere Hospital Quality Incentive Demonstration (HQID) tracked the outcomes of (1) acute myocardial infarction (AMI), (2) congestive heart failure (CHF), (3) pneumonia, (4) coronary artery bypass graft surgery, (5) hip and knee replacement surgery, and (6) perioperative management of surgical patients (including prevention of surgical site infections) through 2009 [14]. Incentivizing certain behaviors in hospitals to provide the highest quality of patient care and outcomes requires that measures of quality be very specific. Therefore, in 2005, CMS began utilizing a website (Hospital Compare Web) that allowed public access to data on hospital performance measures, including resource utilization, spending, clinical outcomes, and process quality. The VBP established a baseline level of performance for each hospital to compare outcomes to other hospitals (for a specific year) and to itself (for previous years). Establishing a baseline is important because errors at the beginning of a process improvement translate to problems in later time frames [15]. Random variation during the time a baseline is being set could show either higher or lower quality when compared to a steady state of the organization. In such cases, the improvement (or lack thereof) may be caused by this random error as the process improvement continues. Notably, these measures of outcomes may not be related to the existing processes. Likewise, subsequent morale issues among the employees may cause them to become disinterested in the process. The VBP uses larger sample sizes to avert such problems; however, not all process improvement measures have high reliability. These variations may be outgrowths from procedures that rarely occur in some hospitals. Thus, longer term use of VBP will require that CMS balance statistical reliability in the data while increasing the number of hospitals in the program. Adjustments may need to be made as the overall program progresses [16].

In 2016, approximately 60% of hospitals received total penalties or incentives of less than±$30,000 [17]. The maximum VBP penalties or financial bonuses were capped at+/-2% for the hospitals. Their results also indicated that the cumulative financial effects can significantly impact hospitals if the same institutions are receiving penalties each year. Carroll and Clement [16] studied the impact of the VBP program on the hospital’s organizational, community, and management characteristics. They found that 24% of the hospitals improved their VBP program performance. However, 11% of hospitals were being penalized each year. Results also showed that 18% of the hospitals performed well in the early years but experienced declines later. Thus, the contributions to the VBP program scores were associated with the different types of hospital characteristics. Many studies have demonstrated that VBP financial incentives positively impacts quality of care and patient safety [18–21]. Lee et al. [22] noted that the penalized hospitals changed their operations to reflect operational performance, thereby resulting in improved VBP metrics to enhance hospital operational outcomes and financial revenues. However, other studies have found that pay-for-performance programs such as VBP generate little or no improvements in patient outcomes [23–26]. Notably, some hospitals which participated in VBP were not impacted during the initial VBP implementation period [27]. As an example, one component of the VBP measurements is mortality rate; however, Figueroa et al. [15] indicated that VBP did not improve patient mortality rates from 2008 to 2013. Likewise, Mendelson et al. [28] found that VBP did not improve hospital cost efficiency.

2.2Hospital consumer assessment of healthcare providers and systems (HCAHPS)

The Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) patient satisfaction survey is an instrument developed by the Centers for Medicare and Medicaid Services (CMS) and the Agency for Healthcare Research and Quality (AHRQ) to establish a standardized process for reporting patients’ perceptions of care. HCAHPS results have remained publicly available since 2008 [39]. A secondary outcome of the HCAHPS survey encompasses various domains, including nursing and physician communication, pain control, environmental factors, medication communication, quality of discharge information, as well as patient perceptions on the responsiveness of staff. For an institution to be considered a high performer in these domains, the hospital must score above the median in top-box responses for each survey question [40]. As described above, HCAHPS surveys are a component of VBP’s quality measurements which consist of patient experiences with regard to hospital care. Through the integration of HCAHPS scores into the value-based incentive payments, CMS has placed greater importance on patient satisfaction as an aspect of quality of care. The HCAHPS survey, in conjunction with changes in Medicare’s reimbursement practices, has placed great value on the patient experience. These changes in Medicare reimbursement policies tend to modify hospitals’ strategic management plans as they attempt to secure peak levels of reimbursement. In the past, successful Medicare policies were often replicated by the private sector within only a few years. Now, hospitals must learn to navigate these policy changes before private insurers implement similar policies [7]. While it is clear that delivering patient-centered care should be a priority for health care providers, it is not evident that patient satisfaction presents a direct link to patient safety or effectiveness of care. Various studies with the aim of developing direct associations among patient safety, satisfaction, and outcomes have failed to produce consistent results [39].

In 2013, all acute care hospitals in the United States (excluding those in Maryland) were eligible to participate in the VBP program. Exceptions to this opportunity were psychiatric hospitals, rehabilitation hospitals, long-term care facilities, children’s hospitals, and cancer hospitals. CMS had revised the payment process to encourage these facilities to improve clinical processes of care and patient experience of care. As such, hospitals were required to report at least ten cases each in at least four of twelve measures included in the clinical processes of care domain, and/or submit at least 100 completed HCAHPS. The HCAHPS survey serves as the standard means to measure satisfaction of patients and how they perceive hospital care. This instrument measures: 1) communication with nurses, 2) communication with doctors, 3) responsiveness of hospital staff, 4) pain management, 5) communication about medicines, 6) cleanliness and quietness of the hospital environment, 7) discharge information, and 8) overall rating of the hospital. Hospitals saw even higher financial incentives (or penalties) as the VBP program evolved based on their delivery of higher value care compared to previous incentives based on higher volume of patients and procedures. Due to such constraints, hospital policy makers and doctors should consider the various challenges when considering the merits of VBP in providing generally higher quality care to patients [14]. In 2015, HAC was added as part of the VBP program by CMS to reduce payments by 1% for hospitals when ranked above the 75th percentile of risk-adjusted total HAC scores. The HAC penalty emphasizes the necessity for hospitals to remain quality-focused and safety-focused, which could potentially influence their performance on patient satisfaction with care, another crucial performance metric tied to reimbursement.

One study by Menendez and Ring [31] assessed the correlation between HAC scores and HCAHPS patients’ perspectives of care measurements. They found that higher quality of care (which corresponds to a lower HAC score) is associated with a better patient experience (reflected by higher HCAHPS scores). Meanwhile, Silvera [41] stated that hospital size influences the relationship between patient safety patient experiences. With larger hospitals, the providers’ ability to deliver care might be less responsive to individual needs, thereby increasing the adverse events on patient safety. Tzeng et al. [42] provided a brief study of the relationship between the rate of pressure ulcers and in-patient satisfaction levels. Higher rates of pressure ulcers were associated with lower in-patient satisfaction levels, especially in staff responsiveness and nurses’ communication. A more in-depth study with Odom-Maryon et al. [12] explored the occurrence of hospital-acquired pressure ulcers with HCAHPS scores in Medicare patients 65 years or older. Their findings revealed that pressure ulcers occurred less frequently among patients who indicated higher patient satisfaction HCAHPS scores for quietness at night and satisfactory nurse communication. The majority of studies indicate that hospitals which provide better quality care usually receive higher patient satisfaction responses. Thus, the current study seeks to determine whether the VBP program actually reflects quality of care and patient responsiveness via HCAHPS regarding the care that they received. As such, the second hypothesis is presented as follows:

Hypothesis 2: A higher Hospital Acquired Condition (HAC) Patient Safety score is associated with lower HCAHPS.

It is intuitive that doctors who communicate well develop better relationships with their patients; likewise, better relationships tend to improve patient ratings of the quality of service for the hospital. Hospital administrators should stress open and clear communication between doctors and patients in order to avoid problems ranging from misdiagnoses to incorrectly following treatment plans. In one study, 75% of the orthopedic surgeons indicated that they communicated satisfactorily with their patients, but only 21% of the patients reported satisfactory communication with their doctors. This represents a dramatic difference in the perceptions of good communication between physicians and patients [43]. The actions of physicians and nurses consist of the largest element of the HCAHPS survey scores due to their extensive interaction with patients. Particularly, caring attitudes and good communication skills are strong predictors of trust and confidence in healthcare providers [44]. Therefore, the quality of communication by medical professionals serves as a crucial measure of perceived patient experiences in the HCAHPS survey.

Carter et al. [45] evaluated patient satisfaction/experience aspects before patients were discharged. In their study, patients were less likely to return to the hospital for further treatment if they were “very satisfied” with the overall care, and the physicians listened to them during their hospital stay. Hence, identifying patient demands prior to discharge may improve overall satisfaction scores. Cruz et al. [46] found that communication, teamwork, and good leadership are essential reasons for a well-designed discharge plan intervention. As a result, good discharge planning is associated with higher HCAHPS survey results. Likewise, patients with proper discharge plans were more likely to recommend hospitals where they had received treatment. In an earlier study, Merlino et al. [13] demonstrated that when treating patients with more severe illnesses, hospitals frequently obtained lower HCAHPS scores which possibly led to lower reimbursements from CMS. More recently, Jindal et al. [47] studied factors that influenced hospital penalties due to lower quality outcomes. They observed that a hospital was about 40% less likely to be penalized when a higher percentage of patients showed that their doctor communicated well and indicated that staff offered necessary information regarding discharge. Based on the above literature, higher quality care in hospitals is typically associated with higher scores on patient survey responses. Several studies have assessed the relationships regarding patient responses of the quality care and hospital incentives, as well as reimbursement impact. For the present study, patient responses to the quality of care surveys (HCAHPS) were tested to identify whether a relationship exists between HCAHPS scores and a hospital’s reimbursement from CMS. Thus, the third hypothesis is presented as follows:

Hypothesis 3: A higher HCAHPS score is associated with higher Reimbursements.

3.1Data sources and characteristics

The Centers of Medicare & Medicaid Services (CMS) report various data on the Care Compare website regarding healthcare providers’ operations and quality measurement information on Doctors & Clinicians, Hospitals, Nursing homes, etc. The “Hospitals” data includes US hospitals’ quality ratings and patient satisfaction information on over 4,000 Medicare-certified hospitals. For the present study, various categories of secondary data were retrieved from the CMS “Hospitals” database for the analysis. Specifically, we utilize hospitals’ datasets for 2020 and coordinate the items of the Value-Based Purchasing Program (VBP) in the Patient Safety Indicator/Index, HCAHPS, and Reimbursement information. The Patient Safety Indicator (PSI) provides a patient’s safety indicator (e.g. pressure sores) which is indicated in a ratio format. The HCAHPS survey contains different publicly reported measures asking recently-discharged patients to evaluate various aspects of their hospital experience. The HCAHPS includes seven collective variables (Communication with Nurses, Care Transition, Recommend Hospital, Communication about Medicines, Communication with Doctors, Discharge Information, and Staff Responsiveness), two individual variables, and two global variables within each variable based on one to three questions with scales which range from 1 to 5. For example, the questions are related to how well nurses and doctors communicate with patients, how well the staff members communicate with patients about medicine, how responsive hospital staff are to patient needs, whether the patients would recommend the hospital to others, etc. Reimbursements include Medicare spending per beneficiary and Medicare hospital spending by claim which are listed in ratio and average spending format. Table 1: Definitions and Measurements of variables provides information regarding the variable definitions and survey questions for this study [48].

After eliminating all missing values, the data variables of HCAHPS, VBP (Patient Safety Index –PSI), and Reimbursements were utilized for this present study. The remaining total sample size of 2,297 acute care hospitals was applied for our data analysis. Please see Table 2: Sample hospital demographics and characteristics –frequency data by Region (Midwest, Northeast, South, and West) and Type of Hospital (Government, Not-for-Profit, and For-Profit). The hospitals are subject to the Value-Based Purchasing (VBP) Program which withholds specific reimbursement amounts due to low scores on relevant quality measurements. For this study, the control variables were the hospital’s ownership (i.e., Government, Not-For-Profit, For-Profit) and categorization of VBP subjects (i.e. CMS withholds some percentage of the reimbursement: Yes, No). Next, we analyzed the data by applying exploratory factor analysis (EFA) for data reduction to determine the number of constructs with the cutoff point set at 0.4. A total of sixteen (16) items were included in the EFA analysis to stipulate the number and patterns of common factors: Patient Safety (four items), HCAHPS (nine items), and Reimbursements (three items). The output provided three interpreted constructs, along with two or more items for each construct for a total of twelve (12) items to comprise the sample structure: Patient Safety (3), HCAHPS (7), and Reimbursements (2).

3.2Reliability and validity analysis

Reliability is assessed via Cronbach’s coefficient alpha and composite reliability (CR) on three scale constructs. The general principle for Cronbach’s alpha is specified as at least 0.6 as the lower limit for reliability [49]. In this present study, the Cronbach’s alpha (CA) values for Patient Safety, HCAHPS, and Reimbursements are 0.724, 0.812, and 0.922, respectively. Factor analysis was conducted to reduce item responses to a particular score for each of the three construct dimensions. Principle component analysis is used to summarize the original data into a range of scores. All composite reliability values are between 0.834 and 0.938, which indicates acceptable reliability as these values exceed 0.70. Please see Table 3: Measurement scales and Loadings.

The Variance Inflation Factor (VIF) is calculated to assess the presence of multicollinearity among the predictor variables: HCAHPS and Reimbursements. If the VIF is higher than 10, the possibility of multicollinearity would cause a concern. The VIF values in this study reveal that all indicators are less than 7, indicating that no collinearity issues exist. All average variance extracted (AVE) values range between 0.629 and 0.841 (at the construct level) which is greater than 0.5. This implies that convergent validity at the indicator and construct levels is validated. The square root of each AVE is examined for discriminant validity and should be greater than 0.7 [50] and exceed the related inter-construct correlations for reflective constructs. All the square roots of AVE are greater than the related inter-construct correlations. Heterotrait-monotrait (HTMT) ratio of correlation is further conducted to check for discriminant validity based on the variance-based view. Table 4: HTMT provides the specific construct values. All of the HTMT values (ranging from 0.112 to 0.434) from the comparison constructs are less than HTMT_0.85 criterion values which indicates that no discriminant validity issues exist among the latent constructs. Therefore, discriminant validity is confirmed.

3.3Hypothesis tests and research results

For this present study, the research model (Fig. 1) is implemented and analyzed to assess causal relationships in the healthcare industry, especially in hospital settings using Partial Least Squares (PLS) modeling via Smart-PLS. PLS was implemented instead of covariance-based SEM because of PLS having no strict normality distribution assumptions for the data, and both formative and reflective constructs can be modeled within the same study [51]. Table 5 provides R², f², and Q² output. The recommended value for R² is that it should be higher than 10 percent [52]. The R² value for endogenous variables HCAHPS and Reimbursement were 0.072 and 0.157, respectively. The value of f² is an indicator of the effect size. The f² values of 0.02, 0.15, and 0.35 are considered small, medium, ad large effects, respectively [53]. The f² effect size for the predictive value of Patient Safety on Reimbursement and HCAHPS are 0.003 and 0.011, respectively, which shows that Patient Safety has a small effect in generating the R² for Reimbursement and HCAHPS. By estimating the scale of the R² value as a measure of predictive accuracy, Q² values indicate the model’s predictive relevance if Q² is greater than 0 [54]. Both Q² values for HCAHPS and Reimbursement (0.069 and 0.044, respectively) are less than 0 which indicates that the model has rather low predictive relevance for these constructs. The bootstrap procedure of resampling (1,000 subsamples) is utilized to acquire the path coefficients, t-statistics, and the 95% confidence interval. The input for the partial least squares model estimations is based on scores of a total of three dimensions: Patient Safety, HCAHPS, and Reimbursements. The results for the present hypotheses are shown in Table 6: Summary for Hypothesis tests which lists the path coefficients, t-statistics, and 95% confidence interval.

Hypothesis 1 states that a higher Patient Safety score is associated with lower Reimbursements which would indicate a negative relationship. Based on the PLS-SEM output from Table 6, Patient Safety score is significantly related to Reimbursements (p < 0.01). However, Patient Safety shows a significant positive relationship (p < 0.01) with Reimbursement which indicates the opposite relationship expected in Hypothesis 1. This finding suggests that in spite of a higher patient safety index in serious complications, blood clots, and pressure sore rates, the hospital typically receives higher reimbursements for these patients with higher complication rates.

Hypothesis 2 states that a higher Patient Safety score is related to lower HCAHPS which indicates a negative relationship. The results from Table 6 reveal that Patient Safety and HCAHPS had a significant negative relationship (p < 0.01). This outcome provided the same results as originally stated in Hypothesis 2 because a lower score in Patient Safety ratings (indicating fewer complications) is related to higher patient satisfaction. According to the CMS website, lower hospital patient safety indexes are indicative of higher quality care. As such, this outcome implies that a lower patient safety rating for patients relates to a higher consumer (patient) assessment based on the consumer/patient ratings regarding doctors, nurses, medicine, staff, discharge procedures, care transition, and hospital recommendations.

Hypothesis 3 states that a higher HCAHPS score is positively related to higher Reimbursements. Based on the results presented in Table 6, a facility’s HCAHPS score is significantly related to Reimbursements (p < 0.01). However, HCAHPS and Reimbursements revealed a significant negative relationship (p < 0.01) which implies the opposite relationship predicted in Hypothesis 3. This finding suggests that although a hospital receives a lower rating on patient assessments, it may still obtain larger reimbursements from Medicare patient services.

For the current research model, two (2) control variables are utilized: Hospital Type and HAC subjects of withholding of reimbursements. The control variable— hospital ownership (Government, Not-For-Profit, For-Profit)— displays varied results for the dependent variables of Reimbursements and HCAHPS. Reimbursements is positively linked to the type of ownership which implies that For-Profit hospitals tend to obtain higher reimbursements from patient services compared to Government or Not-For-Profit hospitals. Additionally, For-Profit hospitals received lower HCAHPS compared to Government or Not-For-Profit hospitals. The control variable –HAC Subjects (Yes, No) withholding hospitals –provided different results for dependent variables of Reimbursements and HCAHPS. The hospitals subject to withholding payments still obtained higher reimbursements compared to the hospitals without being subject to withholding penalties. Nevertheless, the hospitals were subjected to CMS penalties of withholding payments due to lower HCAHPS scores.

4.1Theoretical implications

Research question 1 predicted that a lower HAC score (i.e. better patient safety index ratings) would be associated with higher reimbursements. The actual results indicated that high “Patient Safety” (high HAC scores) were associated with higher “Reimbursements” for patient services, in contrast to the research hypothesis which predicted a negative relationship. These results are in alignment with the findings of Rajaram et al. [36] and Zuckerman et al. [35] which indicated that VBP incentive policies did not considerably increase reimbursements for the hospitals. Their findings indicated that hospitals with better performance on process care were still penalized more frequently. Thus, hospitals with higher HAC scores (due to pressure sores, for example) could still receive greater reimbursements. Furthermore, hospitals obtained higher payment amounts per episode of hospital admittance for Medicare beneficiaries even though the HAC scores remained high. For research question 2, does a low HAC score (i.e. better patient safety index ratings) lead to higher HCAHPS ratings? The results of this present study reveal that better “Patient Safety” index ratings (i.e. low HAC scores) are associated with higher HCAHPS ratings. These findings were similar to the results of several other studies [12, 31, 41, 42] noting that better patient safety is correlated with patient satisfaction. With regard to research question 3, do HCAHPS ratings affect reimbursement payment amounts? The results of this present study suggest that HCAHPS ratings were negatively related to reimbursement amounts. Although the patient/consumer gave lower HCAHPS scores, the hospitals still collected higher reimbursement payments from CMS. Thus, the results of this present study were in opposition to Hypothesis 3 which predicted the relationship would be positive, as indicated by the literature review. The above theoretical results contributed some evidence that the VBP program is not achieving its intended goal of penalizing hospitals with lower quality patient care. Indeed, when hospitals report lower patient quality of care and lower patient satisfaction, the hospital generally did not receive any financial penalty (i.e. decreased reimbursements). However, the results of this present study confirmed that lower patient quality care was associated with lower patient satisfaction.

4.2Practical implications

The findings of this present study indicate that data obtained from the CMS Care Compare website for the acute-care hospitals accurately reflect quality outcomes. Nevertheless, the VBP was not associated with financial incentives or penalties when the hospitals reported high patient safety issues. The VBP was designed to provide financial incentives to hospitals which consistently provide cost-effective, high-quality care for patients. However, the effectiveness of value-based reimbursement and the concept of continuous improvement is constrained due to a lack of unified measurement objectives across healthcare platforms. This process strongly enhances the concepts of hospital quality that impact patient outcomes. Meanwhile, other studies question the associations between improvements in quality and hospital reimbursements [15, 24, 27]. Additionally, incentivizing certain criteria from HCAHPS may produce unintended consequences. Attempts to improve scores on the HCAHPS criteria may cause hospitals to divert resources from certain activities that have yet to be emphasized by VBP. The stated goals for VBP (i.e. improving patient outcomes efficiently) may not reduce overall patient morbidity and mortality, nor improve patient satisfaction at lower costs. Nevertheless, by placing more power in the hands of patients within the US healthcare system, results of the HCAHPS survey have become a determining factor in the reward system of hospital payments. Consequently, the information obtained from patients on the HCAHPS survey directly impacts provider revenue. Since such policies are expected to remain in effect for the foreseeable future, healthcare administrators should develop processes and procedures to meet these demands.

4.3Limitations and future research

Several important limitations have been noted for this present study. First, several options exist to improve the VBP program. Rather than imposing a strict penalty for hospitals in the highest quartile, CMS should give progressively lower reimbursements for hospitals in the highest quartile of performance issues as measured by performance and patient satisfaction criteria. The Hospital-Acquired Condition (HAC) Reduction Program uses such a system, and it is more equitable. Second, legitimate concerns have been raised regarding comparisons between hospitals which serve more disparate populations as compared to competing providers. Hence, outcome measures should better represent local patient demographics. The suggestion would be to use a graduated scale for reduced reimbursement based on the percentage of disadvantaged patients that the hospital serves. This would protect hospitals which treat patients who have higher rates of chronic health issues, along with patients who cannot afford to pay for services rendered. Finally, this present study implies that the quality of healthcare is perceived not only from patients’ opinions of experience with physicians, nurses, or the environment, but also from the quality of outcome performance, such as patient safety in HAC scores. Notably, the findings of this present study show that a poor patient safety index is not related to lower reimbursement payments. Therefore, these results indicate weaknesses in the VBP’s policy of financially penalizing hospitals for poor patient safety ratings. Hence, existing quality measures in the healthcare industry may not completely and accurately reflect the performance of healthcare organizations and providers.

Future researchers may wish to consider the following suggestions. This present study was based on a single year of data. More extensive longitudinal studies should consider the patient safety index compared to payments over a minimum period of three years. Additionally, further research could determine whether the same hospitals with patient safety index ratings above the Medicare 75th percentile for three consecutive years, for example, still obtain the full amount of reimbursement without financial penalty. In order to create legitimate comparisons that are valid across all health care environments, stakeholders must reach a unified stance on the definition and assessment of quality. The VBP program includes five collective measurements (1. Mortality and complications, 2. Healthcare-associated infections, 3. Patient safety, 4. Patient experience, and 5. Efficiency and cost reduction). The present study primarily investigated patient safety and patient experiences. A full-scale study might consider including all five (5) collective measurements by comparing the impact of reimbursements on overall hospital quality.