Risk-standardized Complication Rate (RSCR) following Elective Primary Total Hip Arthroplasty (THA) and/or Total Knee Arthroplasty (TKA) for Eligible Clinicians and Eligible Clinician Groups

Risk variables for this THA/TKA Complications measure for ECs and EC Groups were derived from the development of the original Hospital-level THA/TKA Complications measure (CBE #1550; FFS-only version).  We describe the original selection of variables below.

Our goal in selecting risk factors for adjustment was to develop parsimonious models that included clinically relevant variables strongly associated with the risk of complication in the 90 days following an index procedure. We used a two-stage approach, first identifying the comorbidity or clinical status risk factors that were most important in predicting the outcome, then considering the potential addition of social risk factors.

The original Hospital-level THA/TKA Complications measure was developed with ICD-9. When ICD-10 became effective in 2015, we transitioned the measure to use ICD-10 codes as well. ICD-10 codes were identified using 2015 General Equivalence Mappings (GEM) mapping software. We then enlisted the help of clinicians with expertise in relevant areas to select and evaluate which ICD-10 codes map to the ICD-9 codes used to define this measure during development. A code set is attached in field 1.13a. (Data Dictionary).

For risk model development, we started with Condition Categories (CCs) which are part of CMS’s Hierarchical Condition Categories (HCCs). The HCC system groups the 70,000+ ICD-10-CM and 17,000+ ICD-9-CM codes into larger clinically coherent groups (201 CCs) that are used in models to predict mortality or other outcomes (Pope et al. 2001; 2011). The HCC system groups ICD codes into larger groups that are used in models to predict medical care utilization, mortality, or other related measures.

To select candidate variables, a team of clinicians reviewed all CCs and excluded those that were not relevant to the Medicare population or that were not clinically relevant to the complication outcome (for example, attention deficit disorder, female infertility). All potentially clinically relevant CCs were included as candidate variables and, consistent with CMS’s other claims-based measures, some of those CCs were then combined into clinically coherent CC groupings.

To inform final variable selection, a modified approach to stepwise logistic regression was performed. The Development Sample was used to create 1,000 “bootstrap” samples. For each sample, we ran a logistic stepwise regression that included the candidate variables. The results were summarized to show the percentage of times that each of the candidate variables was significantly associated with mortality (p<0.01) in each of the 1,000 repeated samples (for example, 90 percent would mean that the candidate variable was selected as significant at p<0.01 in 90 percent of the times). We also assessed the direction and magnitude of the regression coefficients.

The clinical team reviewed these results and decided to retain risk adjustment variables above a predetermined cutoff, because they demonstrated a strong and stable association with risk of complication and were clinically relevant. Additionally, specific variables with particular clinical relevance to the risk of complications were forced into the model (regardless of percent selection) to ensure appropriate risk adjustment for THA/TKA. These included variables representing markers for end of life/frailty, such as:

Markers for end of life/frailty:

• Decubitus Ulcer or Chronic Skin Ulcer (CC 157-CC 161)
• Metastatic and Other Major Cancers (CC 8-CC 12)
• Osteoporosis and Other Bone/Cartilage Disorders (CC 43)
• Chronic Kidney Disease, Stage 5 (CC 136)
• Hemiplegia, Paraplegia, Paralysis, Functional disability (CC 70-CC 74, CC 103, CC 104, CC 189-CC 190)
• Stroke (CC 99-CC 100)

This resulted in a final risk-adjustment model that included 33 variables.

Expanding the measure to include outpatient procedures

After developing the Hospital-level THA/TKA Complications measure for the hospital setting, the measure was re-specified (and CBE endorsed) for the EC/EC Group setting.  Following that, the clinician-level measure was expanded to include procedures performed in inpatient and outpatient settings and was re-specified to include outcomes from outpatient settings. This expanded clinician-level THA/TKA Complications measure, with both inpatient and outpatient procedures (for cohort and outcomes) for ECs and EC Groups is the measure currently under consideration for CBE endorsement with this submission.

Because we expanded the THA/TKA Complications measure for ECs and EC Groups to include inpatient and outpatient procedures, we conducted exploratory analyses to examine the similarities and differences between the inpatient and outpatient cohorts with respect to risk factors used in the existing clinician-level measure, as well as basic non-clinical patient-level characteristics. The existing risk model includes several factors that are strong predictors of patient frailty and functional status, including malnutrition, osteoporosis and vertebral fractures, dementia, paralysis, and decubitus ulcers.  

We examined risk‐adjustment variables for patients’ comorbid conditions identified in both inpatient and outpatient claims for the 12 months prior to the index encounter, as well as those coded as POA for patients in the inpatient cohort (Krumholz et al., 2019). Since POA indicators are not available in outpatient claims, for the outpatient cohort, conditions that may represent adverse outcomes due to care received during the index encounter are not adjusted for in the model. 

Consistent with the currently implemented THA/TKA Complications measure for ECs and EC Groups, the final risk-adjustment model included 34 demographic and clinical comorbidity variables along with risk adjustment for the setting. As expected, the risk factor frequencies tend to be higher among patients in the inpatient setting when compared to outpatient and ASC settings, with a few exceptions.  

This respecified clinician-level measure includes patients undergoing procedures in the inpatient, outpatient and ASC settings. Initial empiric analyses indicated that observed complication rates were lower after hospital outpatient/ASC procedures compared to inpatient procedures despite relatively modest differences in clinical risk factor frequency across settings. We concluded there are multiple factors influencing the decision to perform elective THA/TKA procedures in the inpatient versus outpatient setting. These factors include clinical risk assessment by the surgical team (including underlying patient frailty) hospital policies and resources, patient preference, and social determinants of health (SDOH) such as transportation, access to care, housing situation, home support, health literacy, and income (Silber et al., 2023). We defined the clinical setting as inpatient versus outpatient (observation stay or day surgery) or ASC using administrative encounter information.   

To explore the influence of clinical setting on the risk model, we then assessed the model performance of a single combined model but with indicator variables added to reflect inpatient versus outpatient and ASC settings.  

The odds of THA/TKA complications after surgery are significantly higher in the inpatient setting (OR, 1.39 (95% CI, 1.28, 1.51) and significantly lower in the hospital outpatient setting (OR, 0.90 (95% CI, 0.83, 0.98) compared to the ASC setting (the reference). Most risk factors are associated with higher odds of THA/TKA surgery complications except Other major cancers, Respiratory/heart/digestive/urinary/other neoplasms, and History of COVID.  

Clinical and Social Risk Factor Conceptual Model 

Although some recent literature evaluates the relationship between patient social risk factors and the complication outcome, few studies directly address causal pathways or examine the role of the hospital in these pathways. 

The social risk factors that have been examined in the literature can be categorized into three domains: (1) patient-level variables, (2) neighborhood/community-level variables, and (3) hospital-level variables. Patient-level variables describe the characteristics of individual patients and include the patient’s race and ethnicity, income, or education level. For example, Black and Hispanic patients have been shown to experience higher rates of postoperative complications, longer lengths of stay, and more frequent non-home discharges (Rudisill et al., 2023; Usiskin and Misra, 2022; Brown, Paisner, and Sassoon, 2022)). These disparities are often influenced by socioeconomic factors such as lower income or education levels, which can limit access to healthcare services, preventive care, or rehabilitation (Alvarez et al., 2022). Additionally, barriers such as language proficiency or health literacy may further exacerbate these challenges and contribute to poorer outcomes and higher rates of complications for patients with social risk factors (Suleiman et al., 2021). Neighborhood/community-level variables use information from sources such as the American Community Survey (ACS) as either a proxy for individual patient-level data or to measure environmental factors. Studies using these variables use one-dimensional measures such as median household income or composite measures such as the Agency for Healthcare Research and Quality (AHRQ)-validated Socioeconomic Status (SES) index score (Blum et al., 2014; Courtney et al., 2016; Martsolf et al., 2016; White et al., 2018). Some of these variables may include the local availability of clinical providers (Herrin et al., 2015; Herrin et al., 2016). Hospital-level variables measure attributes of the hospital which may be related to patient risk. Examples of hospital-level variables used in studies are ZIP code characteristics aggregated to the hospital level, or the proportion of Medicaid patients served in the hospital (Gilman et al., 2014; Joynt et al., 2013; Jha et al., 2011; Xu et al., 2017).

The conceptual relationship, or potential causal pathways by which these possible social risk factors influence the risk of complication following an acute illness or major surgery, like the factors themselves, are varied and complex. There are at least four potential pathways that are important to consider:

1. Comorbidities and social risk. Patients who have lower income/education/literacy or unstable housing may have a worse general health status and may present for their hospitalization or procedure with a greater severity of underlying illness. These social risk factors, which are characterized by patient-level or neighborhood/community-level (as proxy for patient-level) variables, may contribute to worse health status at admission due to competing priorities (for example, restrictions based on job), lack of access to care (e.g., geographic, cultural, or financial), or lack of health insurance. Given that these risk factors all lead to worse general health status, this causal pathway should be largely accounted for by current clinical risk adjustment. We note that empirically, patient comorbidities and social risk factors overlap in their contribution to higher risk of the outcome, as shown by our empirical evidence (see Section 5.3) demonstrating the attenuating impact of model variables on the odds ratios for each social risk factor (ADI; DE).
2. Differential care. A second pathway by which social risk factors may contribute to complications risk is that patients may not receive equivalent care within a facility. For example, patients with social risk factors such as lower education may require differentiated care (e.g. provision of lower literacy information – that they do not receive).
3. Low-quality hospitals. Patients with social risk factors may receive care at lower quality hospitals. Patients of lower income, lower education, or unstable housing may not have the same access to high quality facilities, in part, because such facilities may be less likely to be found in geographic areas with large populations of patients with social risk factors (Fahrenbach et. al., 2020). Thus, patients with low income may be more likely to be treated in lower quality hospitals, which may contribute to an increased risk of post-operative complications/readmission. In addition, or alternatively, low-quality hospitals may not implement the evidence-based interventions to reduce the risk of readmission, such as post-discharge follow up; patients with social risk factors are known to have lower rates of follow up after discharge and higher rates of post-discharge acute care.
4. Residual risk. Some social risk factors, such as income or wealth, may affect the likelihood of complications without directly affecting health status at admission or the quality of care received during the hospital stay. For instance, while a hospital may make appropriate care decisions and provide tailored care and education, a lower-income patient may have a worse outcome post-discharge due to competing financial priorities that don’t allow for adequate recuperation or access to needed treatments, or a lack of access to care outside of the hospital.

Based on our previous work to identify social risk factors (“Additional Testing of Gaps in Mortality/Complication Measure Score Performance Among Selected Subpopulations of Interest Defined by Social Risk Factor (SRF) for the 2024 MUC Submissions,” 2024), the social risk factors used for the additional testing and the rationale were:  

1. Dual eligible (DE) status: DE status (i.e., enrolled in both Medicare and Medicaid) for a discharge is derived using the beneficiary enrollment data file in the Integrated Data Repository (IDR). The data includes monthly enrollment status, and a patient is considered DE for an index admission if they are enrolled in both Medicare and Medicaid in the month of discharge date of the admission (Centers for Medicare & Medicaid Services, 2024). We recognize that Medicare-Medicaid dual eligibility has limitations as a proxy for patients' income or assets because it is a dichotomous outcome and does not provide a range of values. However, the dual-eligibility status for over 65-year-old Medicare patients is valuable, as this qualification takes into account both income and assets and is consistently applied across states for the older population.
2. High Area Deprivation Index status (ADI): ADI is a multidimensional measure of socioeconomic status of a geographical area (Petterson et al., 2023). It considers 17 components, comprised of 4 socioeconomic domains, including education, income/employment, housing, household characteristics. It measures at the census block group level and is calculated as a ranking from 0 to 100, with 0 meaning least deprived and 100 meaning most deprived. A census block group is a geographical unit used by the US Census Bureau and is the smallest geographical unit for which the bureau publishes sample data. The target size for block groups is 1,500, with a typical population of 600 to 3,000 people. We dichotomized the ADI rankings to greater than or equal to 85 (High ADI) versus less than 85 (Low ADI) per recommendation by the developer of ADI. For this analysis, we linked ADI at the census block group level to a 9-digit zip code.  

The 17 components are listed below:

• Population aged ≥ 25 y with < 9 y of education, %
• Population aged ≥ 25 y with at least a high school diploma, %
• Employed persons aged ≥ 16 y in white-collar occupations, %
• Median family income, $
• Income disparity
• Median home value, $
• Median gross rent, $
• Median monthly mortgage, $
• Owner-occupied housing units, % (homeownership rate)
• Civilian labor force population aged ≥16 y unemployed, % (unemployment rate)
• Families below the poverty level, %
• Population below 150% of the poverty threshold, %
• Single-parent households with children aged < 18 y, %
• Households without a motor vehicle, %
• Households without a telephone, %
• Occupied housing units without complete plumbing, % (log)
• Households with more than 1 person per room, % (crowding)

References

Alvarez, P. M., McKeon, J. F., Spitzer, A. I., Krueger, C. A., Pigott, M., Li, M., & Vajapey, S. P. (2022). Socioeconomic factors affecting outcomes in total knee and hip arthroplasty: a systematic review on healthcare disparities. Arthroplasty, 4(1), 36.

Blum AB, Egorova NN, Sosunov EA, et al. Impact of socioeconomic status measures on hospital profiling in New York City. Circulation Cardiovascular quality and outcomes 2014; 7:391-7. 

Centers for Medicare & Medicaid Services. (2024). Dually eligible Individuals – categories. https://www.cms.gov/medicare-medicaid-coordination/medicare-and-medicai…

Courtney M, Huddleston J, Iorio R, Markel D. Socioeconomic Risk Adjustment Models for Reimbursement Are Necessary in Primary Total Joint Arthroplasty. July 2016; 32(1):1-5. https://doi.org/10.1016/j.arth.2016.06.050.

Fahrenbach, J., Chin, M. H., Huang, E. S., Springman, M. K., Weber, S. G., & Tung, E. L. (2020). Neighborhood Disadvantage and Hospital Quality Ratings in the Medicare Hospital Compare Program. Medical care, 58(4), 376–383. https://doi.org/10.1097/MLR.0000000000001283

Gilman M, Adams EK, Hockenberry JM, et al. California safety-net hospitals likely to be penalized by ACA value, readmission, and meaningful-use programs. Health Aff (Millwood). Aug 2014; 33(8):1314-22.

Herrin J, Kenward K, Joshi MS, Audet AM, Hines SJ. Assessing Community Quality of Health Care. Health Serv Res. 2016 Feb;51(1):98-116. doi: 10.1111/1475-6773.12322. Epub 2015 Jun 11. PMID: 26096649; PMCID: PMC4722214.

Herrin J, St Andre J, Kenward K, Joshi MS, Audet AM, Hines SC. Community factors and hospital readmission rates. Health Serv Res. 2015 Feb;50(1):20-39. doi: 10.1111/1475-6773.12177. Epub 2014 Apr 9. PMID: 24712374; PMCID: PMC4319869.

Jha AK, Orav EJ, Epstein AM. Low-quality, high-cost hospitals, mainly in South, care for sharply higher shares of elderly black, Hispanic, and medicaid patients. Health affairs 2011; 30:1904-11.

Joynt KE, Jha AK. Characteristics of hospitals receiving penalties under the Hospital Readmissions Reduction Program. JAMA. Jan 23 2013; 309(4):342-3.

Krumholz HM, Coppi AC, Warner F, et al. Comparative effectiveness of new approaches to improve mortality risk models from Medicare claims data. JAMA Network Open. 2019;2(7):e197314-e197314.

Martsolf G, Barrett M, Weiss A, Kandrack R, Washington R, Steiner C, Mehrotra A, SooHoo N, Coffey R. Impact of Race/Ethnicity and Socioeconomic Status on Risk-Adjusted Hospital Readmission Rates Following Hip and Knee Arthroplasty, The Journal of Bone and Joint Surgery. 2016;98(16):1385-1391. https://doi.org/10.2106/JBJS.15.00884.

Petterson S. Deciphering the Neighborhood Atlas Area Deprivation Index: the consequences of not standardizing. Health Aff Sch. 2023 Nov 3;1(5):qxad063. doi: 10.1093/haschl/qxad063. PMID: 38756979; PMCID: PMC10986280. 

Pope GC, Ellis RP, Ash AS, et al. Diagnostic cost group hierarchical condition category models for Medicare risk adjustment. Final Report to the Health Care Financing Administration under Contract Number 500-95-048. 2001; http://www.cms.hhs.gov/Reports/downloads/pope_2000_2.pdf.

Pope GC, Kautter J, Ingber MJ, et al. Evaluation of the CMS-HCC Risk Adjustment Model: Final Report. 2011; https://www.cms.gov/Medicare/HealthPlans/MedicareAdvtgSpecRateStats/dow….

Rudisill, S. S., Varady, N. H., Birir, A., Goodman, S. M., Parks, M. L., & Amen, T. B. (2023). Racial and ethnic disparities in total joint arthroplasty care: A contemporary systematic review and meta-analysis. The Journal of Arthroplasty, 38(1), 171-187.e18. https://doi.org/10.1016/j.arth.2022.08.006

Silber JH, Rosenbaum PR, Reiter JG, Jain S, Ramadan OI, Hill AS, Hashemi S, Kelz RR, Fleisher LA. The Safety of Performing Surgery at Ambulatory Surgery Centers Versus Hospital Outpatient Departments in Older Patients With or Without Multimorbidity. Med Care. 2023 May 1;61(5):328-337. doi:10.1097/MLR.0000000000001836. Epub 2023 Mar 17. PMID: 36929758; PMCID: PMC10079624. 

Suleiman, L. I., Tucker, K., Ihekweazu, U., Huddleston, J. I., & Cohen-Rosenblum, A. R. (2022). Caring for diverse and high-risk patients: Surgeon, health system, and patient integration. The Journal of Arthroplasty, 37(8), 1421–1425. https://doi.org/10.1016/j.arth.2022.04.077

Usiskin, I., & Misra, D. (2022). Racial disparities in elective total joint arthroplasty for osteoarthritis. ACR Open Rheumatology, 4(4), 306–311. https://doi.org/10.1002/acr2.11399

White, R.S., Sastow, D.L., Gaber-Baylis, L.K. et al. Readmission Rates and Diagnoses Following Total Hip Replacement in Relation to Insurance Payer Status, Race and Ethnicity, and Income Status. J. Racial and Ethnic Health Disparities 5, 1202–1214 (2018). https://doi.org/10.1007/s40615-018-0467-0.

Xu HF, White RS, Sastow DL, Andreae MH, Gaber-Baylis LK, Turnbull ZA. Medicaid insurance as primary payer predicts increased mortality after total hip replacement in the state inpatient databases of California, Florida and New York. J Clin Anesth. 2017;43:24‐32. doi:10.1016/j.jclinane.2017.09.008.