ABSTRACT
BACKGROUND: When evaluating the results of clinical research studies, readers need to know that patients perceive effect sizes, not p values. Knowing the minimum clinically important difference (MCID) and the patient-acceptable symptom state (PASS) threshold for patient-reported outcome measures helps us to ascertain whether our interventions result in improvements that are large enough for patients to care about, and whether our treatments alleviate patient symptoms sufficiently. Prior studies have developed the MCID and PASS threshold for the Hip Disability and Osteoarthritis Outcome Score for Joint Replacement (HOOS JR) and Knee Injury and Osteoarthritis Outcome Score for Joint Replacement (KOOS JR) anchored on satisfaction with surgery, but to our knowledge, neither the MCID nor the PASS thresholds for these instruments anchored on a single-item PASS question have been described. QUESTIONS/PURPOSES: (1) What are the MCID (defined here as the HOOS/KOOS JR change score associated with achieving PASS) and PASS threshold for the HOOS JR and KOOS JR anchored on patient responses to the single-item PASS instrument? (2) How do patient demographic factors such as age, gender, and BMI correlate with MCID and PASS thresholds using the single-item PASS instrument? METHODS: Between July 2020 and September 2021, a total of 10,970 patients underwent one primary unilateral THA or TKA and completed at least one of the three surveys (preoperative HOOS or KOOS JR, 1-year postoperative HOOS or KOOS JR, and 1-year postoperative single-item anchor) at one large, academic medical center. Of those, only patients with data for all three surveys were eligible, leaving 13% (1465 total; 783 THAs and 682 TKAs) for analysis. Despite this low percentage, the overall sample size was large, and there was little difference between completers and noncompleters in terms of demographics or baseline patient-reported outcome measure scores. Patients undergoing bilateral total joint arthroplasty or revision total joint arthroplasty and those without all three surveys at 1 year of follow-up were excluded. A receiver operating characteristic curve analysis, leveraging a 1-year, single-item PASS (that is, "Do you consider that your current state is satisfactory?" with possible answers of "yes" or "no") as the anchor was then used to establish the MCID and PASS thresholds among the 783 included patients who underwent primary unilateral THA and 682 patients who underwent primary unilateral TKA. We also explored the associations of age at the time of surgery (younger than 65 years or 65 years and older), gender (men or women), BMI (< 30 or ≥ 30 kg/m 2 ), and baseline Patient-Reported Outcome Measure Information System-10 physical and mental component scores (< 50 or ≥ 50) for each of the MCID and PASS thresholds through stratified analyses. RESULTS: For the HOOS JR, the MCID associated with the PASS was 23 (95% CI 18 to 31), with an area under the receiver operating characteristic curve of 0.75, and the PASS threshold was 81 (95% CI 77 to 85), with an area under the receiver operating characteristic curve of 0.81. For the KOOS JR, the MCID was 16 (95% CI 14 to 18), with an area under the receiver operating characteristic curve of 0.75, and the PASS threshold was 71 (95% CI 66 to 73) with an area under the receiver operating characteristic curve of 0.84. Stratified analyses indicated higher change scores and PASS threshold for younger men undergoing THA and higher PASS thresholds for older women undergoing TKA. CONCLUSION: Here, we demonstrated the utility of a single patient-centered anchor question, raising the question as to whether simply collecting a postoperative PASS is an easier way to measure success than collecting preoperative and postoperative patient-reported outcome measures and then calculating MCIDs and the substantial clinical benefit. LEVEL OF EVIDENCE: Level III, therapeutic study.
Subject(s)
Arthroplasty, Replacement, Hip , Knee Injuries , Osteoarthritis , Male , Humans , Female , Aged , Treatment Outcome , Arthroplasty, Replacement, Hip/adverse effects , Patient Reported Outcome Measures , Minimal Clinically Important DifferenceABSTRACT
Primary osteoporosis is characterized by decreasing bone mass and density and reduced bone strength that leads to a higher risk for fracture, especially hip and spine fractures. The prevalence of osteoporosis in the United States is estimated at 12.6% for adults older than 50 years. Although it is most frequently diagnosed in White and Asian females, it still affects males and females of all ethnicities. Osteoporosis is considered a major health issue, which has prompted the development and use of several performance measures to assess and improve the effectiveness of screening, diagnosis, and treatment. These performance measures are often used in accountability, public reporting, and/or payment programs. However, the reliability, validity, evidence, attribution, and meaningfulness of performance measures have been questioned. The purpose of this paper is to present a review of current performance measures on osteoporosis and inform physicians, payers, and policymakers in their selection of performance measures for this condition. The Performance Measurement Committee identified 6 osteoporosis performance measures relevant to internal medicine physicians, only 1 of which was found valid at all levels of attribution. This paper also proposes a performance measure concept to address a performance gap for the initial approach to therapy for patients with a new diagnosis of osteoporosis based on the current American College of Physicians guideline.
Subject(s)
Fractures, Bone , Osteoporosis , Male , Female , Humans , Adult , United States/epidemiology , Quality Indicators, Health Care , Reproducibility of Results , Osteoporosis/diagnosis , Osteoporosis/therapy , Bone Density , Fractures, Bone/epidemiologyABSTRACT
There has been an exponential growth in the use of telemedicine services to provide clinical care, accelerated by the COVID-19 pandemic. Clinical care delivered via telemedicine has become a major and accepted method of health care delivery for many patients. There is an urgent need to understand quality of care in the telemedicine environment. This American College of Physicians position paper presents 6 recommendations to ensure the appropriate use of performance measures to evaluate quality of clinical care provided in the telemedicine environment.
Subject(s)
COVID-19 , Physicians , Telemedicine , Humans , Pandemics , Telemedicine/methods , Delivery of Health CareABSTRACT
BACKGROUND: Unplanned hospital readmissions after total joint arthroplasty (TJA) represent potentially serious adverse events and remain a critical measure of hospital quality. Predicting the risk of readmission after TJA may provide patients and clinicians with valuable information for preoperative decision-making. QUESTIONS/PURPOSES: (1) Can nonlinear machine-learning models integrating preoperatively available patient, surgeon, hospital, and county-level information predict 30-day unplanned hospital readmissions in a large cohort of nationwide Medicare beneficiaries undergoing TJA? (2) Which predictors are the most important in predicting 30-day unplanned hospital readmissions? (3) What specific information regarding population-level associations can we obtain from interpreting partial dependency plots (plots describing, given our modeling choice, the potentially nonlinear shape of associations between predictors and readmissions) of the most important predictors of 30-day readmission? METHODS: National Medicare claims data (chosen because this database represents a large proportion of patients undergoing TJA annually) were analyzed for patients undergoing inpatient TJA between October 2016 and September 2018. A total of 679,041 TJAs (239,391 THAs [61.3% women, 91.9% White, 52.6% between 70 and 79 years old] and 439,650 TKAs [63.3% women, 90% White, 55.2% between 70 and 79 years old]) were included. Model features included demographics, county-level social determinants of health, prior-year (365-day) hospital and surgeon TJA procedure volumes, and clinical classification software-refined diagnosis and procedure categories summarizing each patient's Medicare claims 365 days before TJA. Machine-learning models, namely generalized additive models with pairwise interactions (prediction models consisting of both univariate predictions and pairwise interaction terms that allow for nonlinear effects), were trained and evaluated for predictive performance using area under the receiver operating characteristic (AUROC; 1.0 = perfect discrimination, 0.5 = no better than random chance) and precision-recall curves (AUPRC; equivalent to the average positive predictive value, which does not give credit for guessing "no readmission" when this is true most of the time, interpretable relative to the base rate of readmissions) on two holdout samples. All admissions (except the last 2 months' worth) were collected and split randomly 80%/20%. The training cohort was formed with the random 80% sample, which was downsampled (so it included all readmissions and a random, equal number of nonreadmissions). The random 20% sample served as the first test cohort ("random holdout"). The last 2 months of admissions (originally held aside) served as the second test cohort ("2-month holdout"). Finally, feature importances (the degree to which each variable contributed to the predictions) and partial dependency plots were investigated to answer the second and third research questions. RESULTS: For the random holdout sample, model performance values in terms of AUROC and AUPRC were 0.65 and 0.087, respectively, for THA and 0.66 and 0.077, respectively, for TKA. For the 2-month holdout sample, these numbers were 0.66 and 0.087 and 0.65 and 0.075. Thus, our nonlinear models incorporating a wide variety of preoperative features from Medicare claims data could not well-predict the individual likelihood of readmissions (that is, the models performed poorly and are not appropriate for clinical use). The most predictive features (in terms of mean absolute scores) and their partial dependency graphs still confer information about population-level associations with increased risk of readmission, namely with older patient age, low prior 365-day surgeon and hospital TJA procedure volumes, being a man, patient history of cardiac diagnoses and lack of oncologic diagnoses, and higher county-level rates of hospitalizations for ambulatory-care sensitive conditions. Further inspection of partial dependency plots revealed nonlinear population-level associations specifically for surgeon and hospital procedure volumes. The readmission risk for THA and TKA decreased as surgeons performed more procedures in the prior 365 days, up to approximately 75 TJAs (odds ratio [OR] = 1.2 for TKA and 1.3 for THA), but no further risk reduction was observed for higher annual surgeon procedure volumes. For THA, the readmission risk decreased as hospitals performed more procedures, up to approximately 600 TJAs (OR = 1.2), but no further risk reduction was observed for higher annual hospital procedure volumes. CONCLUSION: A large dataset of Medicare claims and machine learning were inadequate to provide a clinically useful individual prediction model for 30-day unplanned readmissions after TKA or THA, suggesting that other factors that are not routinely collected in claims databases are needed for predicting readmissions. Nonlinear population-level associations between low surgeon and hospital procedure volumes and increased readmission risk were identified, including specific volume thresholds above which the readmission risk no longer decreases, which may still be indirectly clinically useful in guiding policy as well as patient decision-making when selecting a hospital or surgeon for treatment. LEVEL OF EVIDENCE: Level III, therapeutic study.
Subject(s)
Arthroplasty, Replacement, Hip , Arthroplasty, Replacement, Knee , Male , Humans , Female , Aged , United States , Arthroplasty, Replacement, Hip/adverse effects , Patient Readmission , Medicare , Arthroplasty, Replacement, Knee/adverse effects , Machine Learning , Risk Factors , Retrospective StudiesABSTRACT
BACKGROUND: Although patient-reported outcome measures (PROMs) have become a regularly used metric, there is little consensus on the methodology used to determine clinically relevant postoperative outcomes. We systematically reviewed the literature for studies that have identified metrics of clinical efficacy after total hip arthroplasty (THA) including minimal clinically important difference (MCID), patient acceptable symptom state (PASS), minimal detectable change (MDC), and substantial clinical benefit (SCB). METHODS: A systematic review examining quantitative metrics for assessing clinical improvement with PROMs following THA was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines using the MEDLINE database from 2008 to 2020. Inclusion criteria included full texts, English language, primary THA with minimum 1-year follow-up, use of metrics for assessing clinical outcomes with PROMs, and primary derivations of those metrics. Sixteen studies (24,487 THA patients) met inclusion criteria and 11 different PROMs were reported. RESULTS: MCIDs were calculated using distribution methods in 7 studies (44%), anchor methods in 2 studies (13%), and both methods in 2 studies (13%). MDC was calculated in 2 studies, PASS was reported in 1 study using anchor-based method, and SCB was calculated in 1 study using anchor-based method. CONCLUSION: There is a lack of consistency in the literature regarding the use and interpretation of PROMs to assess patient satisfaction. MCID was the most frequently reported measure, while MDC, SCB, and PASS were used relatively infrequently. Method of derivation varied based on the PROM used; distribution method was more frequently used for MCID.
Subject(s)
Arthroplasty, Replacement, Hip , Humans , Benchmarking , Treatment Outcome , Patient Satisfaction , Minimal Clinically Important Difference , Patient Reported Outcome MeasuresABSTRACT
BACKGROUND: Decisions regarding care for osteoarthritis involve physicians helping patients understand likely benefits and harms of treatment. Little work has directly compared patient and surgeon risk-taking attitudes, which may help inform strategies for shared decision-making and improve patient satisfaction. METHODS: We surveyed patients contemplating total joint arthroplasty visiting a high-volume specialty hospital regarding general questions about risk-taking, as well as willingness to undergo surgery under hypothetical likelihoods of moderate improvement and complications. We compared responses from surgeons answering similar questions about willingness to recommend surgery. RESULTS: Altogether 82% (162/197) of patients responded, as did 65% (30/46) of joint replacement surgeons. Mean age among patients was 66.4 years; 58% were female. Surgeons averaged 399 surgeries in 2019. Responses were similar between groups for general, health, career, financial, and sports/leisure risk-taking (P > .20); surgeons were marginally more risk-taking in driving (P = .05). For willingness to have or recommend surgery, as the chance of benefit decreased, or the chance of harm increased, the percentage willing to have or recommend surgery decreased. Between a 70% and 95% chance of moderate improvement (for a 2% complication risk), as well as between a 90% and 95% chance of moderate improvement (for 4% and 6% complication risks), the percentage willing to have or recommend surgery was indistinguishable between patients and surgeons. However, for lower likelihoods of improvement, a higher percentage of patients were willing to undergo surgery than surgeons recommended. Patients were also more often indifferent between complication risks. CONCLUSION: Although patients and surgeons were often willing to have or recommend joint replacement surgery at similar rates, they diverged for lower-benefit higher-harm scenarios.
Subject(s)
Arthroplasty, Replacement, Hip , Arthroplasty, Replacement, Knee , Surgeons , Aged , Female , Humans , Male , Risk-Taking , Surveys and QuestionnairesABSTRACT
BACKGROUND: Absenteeism is costly, yet evidence suggests that presenteeism-illness-related reduced productivity at work-is costlier. We quantified employed patients' presenteeism and absenteeism before and after total joint arthroplasty (TJA). METHODS: We measured presenteeism (0-100 scale, 100 full performance) and absenteeism using the World Health Organization's Health and Work Performance Questionnaire before and after TJA among a convenience sample of employed patients. We captured detailed information about employment and job characteristics and evaluated how and among whom presenteeism and absenteeism improved. RESULTS: In total, 636 primary, unilateral TJA patients responded to an enrollment email, confirmed employment, and completed a preoperative survey (mean age: 62.1 years, 55.3% women). Full at-work performance was reported by 19.7%. Among 520 (81.8%) who responded to a 1-year follow-up, 473 (91.0%) were still employed, and 461 (88.7%) had resumed working. Among patients reporting at baseline and 1 year, average at-work performance improved from 80.7 to 89.4. A Wilcoxon signed-rank test indicated that postoperative performance was significantly higher than preoperative performance (P < .0001). The percentage of patients who reported full at-work performance increased from 20.9% to 36.8% (delta = 15.9%, 95% confidence interval = [10.0%, 21.9%], P < .0001). Presenteeism gains were concentrated among patients who reported declining work performance leading up to surgery. Average changes in absences were relatively small. Combined, the average monthly value lost by employers to presenteeism declined from 15.3% to 8.3% and to absenteeism from 16.9% to 15.5% (ie, mitigated loss of 8.4% of monthly value). CONCLUSION: Among employed patients before TJA, presenteeism and absenteeism were similarly costly. After, employed patients reported increased performance, concentrated among those with declining performance leading up to surgery.
Subject(s)
Arthroplasty, Replacement, Knee , Presenteeism , Absenteeism , Efficiency , Female , Humans , Male , Middle Aged , Surveys and QuestionnairesABSTRACT
BACKGROUND: Evidence-based international expert consensus regarding anaesthetic practice in hip/knee arthroplasty surgery is needed for improved healthcare outcomes. METHODS: The International Consensus on Anaesthesia-Related Outcomes after Surgery group (ICAROS) systematic review, including randomised controlled and observational studies comparing neuraxial to general anaesthesia regarding major complications, including mortality, cardiac, pulmonary, gastrointestinal, renal, genitourinary, thromboembolic, neurological, infectious, and bleeding complications. Medline, PubMed, Embase, and Cochrane Library including Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, NHS Economic Evaluation Database, from 1946 to May 17, 2018 were queried. Meta-analysis and Grading of Recommendations Assessment, Development and Evaluation approach was utilised to assess evidence quality and to develop recommendations. RESULTS: The analysis of 94 studies revealed that neuraxial anaesthesia was associated with lower odds or no difference in virtually all reported complications, except for urinary retention. Excerpt of complications for neuraxial vs general anaesthesia in hip/knee arthroplasty, respectively: mortality odds ratio (OR): 0.67, 95% confidence interval (CI): 0.57-0.80/OR: 0.83, 95% CI: 0.60-1.15; pulmonary OR: 0.65, 95% CI: 0.52-0.80/OR: 0.69, 95% CI: 0.58-0.81; acute renal failure OR: 0.69, 95% CI: 0.59-0.81/OR: 0.73, 95% CI: 0.65-0.82; deep venous thrombosis OR: 0.52, 95% CI: 0.42-0.65/OR: 0.77, 95% CI: 0.64-0.93; infections OR: 0.73, 95% CI: 0.67-0.79/OR: 0.80, 95% CI: 0.76-0.85; and blood transfusion OR: 0.85, 95% CI: 0.82-0.89/OR: 0.84, 95% CI: 0.82-0.87. CONCLUSIONS: Recommendation: primary neuraxial anaesthesia is preferred for knee arthroplasty, given several positive postoperative outcome benefits; evidence level: low, weak recommendation. RECOMMENDATION: neuraxial anaesthesia is recommended for hip arthroplasty given associated outcome benefits; evidence level: moderate-low, strong recommendation. Based on current evidence, the consensus group recommends neuraxial over general anaesthesia for hip/knee arthroplasty. TRIAL REGISTRY NUMBER: PROSPERO CRD42018099935.
Subject(s)
Anesthesia, Epidural/adverse effects , Anesthesia, General/adverse effects , Anesthesia, Spinal/adverse effects , Arthroplasty, Replacement, Hip/methods , Arthroplasty, Replacement, Knee/methods , Anesthesia, Epidural/mortality , Anesthesia, General/mortality , Anesthesia, Spinal/mortality , Arthroplasty, Replacement, Hip/mortality , Arthroplasty, Replacement, Knee/mortality , Evidence-Based Medicine/methods , Humans , Postoperative Complications/mortality , Randomized Controlled Trials as Topic , Treatment OutcomeABSTRACT
BACKGROUND: Identifying patients at risk of not achieving meaningful gains in long-term postsurgical patient-reported outcome measures (PROMs) is important for improving patient monitoring and facilitating presurgical decision support. Machine learning may help automatically select and weigh many predictors to create models that maximize predictive power. However, these techniques are underused among studies of total joint arthroplasty (TJA) patients, particularly those exploring changes in postsurgical PROMs. QUESTION/PURPOSES: (1) To evaluate whether machine learning algorithms, applied to hospital registry data, could predict patients who would not achieve a minimally clinically important difference (MCID) in four PROMs 2 years after TJA; (2) to explore how predictive ability changes as more information is included in modeling; and (3) to identify which variables drive the predictive power of these models. METHODS: Data from a single, high-volume institution's TJA registry were used for this study. We identified 7239 hip and 6480 knee TJAs between 2007 and 2012, which, for at least one PROM, patients had completed both baseline and 2-year followup surveys (among 19,187 TJAs in our registry and 43,313 total TJAs). In all, 12,203 registry TJAs had valid SF-36 physical component scores (PCS) and mental component scores (MCS) at baseline and 2 years; 7085 and 6205 had valid Hip and Knee Disability and Osteoarthritis Outcome Scores for joint replacement (HOOS JR and KOOS JR scores), respectively. Supervised machine learning refers to a class of algorithms that links a mapping of inputs to an output based on many input-output examples. We trained three of the most popular such algorithms (logistic least absolute shrinkage and selection operator (LASSO), random forest, and linear support vector machine) to predict 2-year postsurgical MCIDs. We incrementally considered predictors available at four time points: (1) before the decision to have surgery, (2) before surgery, (3) before discharge, and (4) immediately after discharge. We evaluated the performance of each model using area under the receiver operating characteristic (AUROC) statistics on a validation sample composed of a random 20% subsample of TJAs excluded from modeling. We also considered abbreviated models that only used baseline PROMs and procedure as predictors (to isolate their predictive power). We further directly evaluated which variables were ranked by each model as most predictive of 2-year MCIDs. RESULTS: The three machine learning algorithms performed in the poor-to-good range for predicting 2-year MCIDs, with AUROCs ranging from 0.60 to 0.89. They performed virtually identically for a given PROM and time point. AUROCs for the logistic LASSO models for predicting SF-36 PCS 2-year MCIDs at the four time points were: 0.69, 0.78, 0.78, and 0.78, respectively; for SF-36 MCS 2-year MCIDs, AUROCs were: 0.63, 0.89, 0.89, and 0.88; for HOOS JR 2-year MCIDs: 0.67, 0.78, 0.77, and 0.77; for KOOS JR 2-year MCIDs: 0.61, 0.75, 0.75, and 0.75. Before-surgery models performed in the fair-to-good range and consistently ranked the associated baseline PROM as among the most important predictors. Abbreviated LASSO models performed worse than the full before-surgery models, though they retained much of the predictive power of the full before-surgery models. CONCLUSIONS: Machine learning has the potential to improve clinical decision-making and patient care by helping to prioritize resources for postsurgical monitoring and informing presurgical discussions of likely outcomes of TJA. Applied to presurgical registry data, such models can predict, with fair-to-good ability, 2-year postsurgical MCIDs. Although we report all parameters of our best-performing models, they cannot simply be applied off-the-shelf without proper testing. Our analyses indicate that machine learning holds much promise for predicting orthopaedic outcomes. LEVEL OF EVIDENCE: Level III, diagnostic study.
Subject(s)
Arthroplasty, Replacement, Hip , Arthroplasty, Replacement, Knee , Machine Learning , Minimal Clinically Important Difference , Aged , Disability Evaluation , Female , Humans , Male , Middle Aged , Patient Reported Outcome Measures , Predictive Value of Tests , Registries , Retrospective StudiesABSTRACT
BACKGROUND: Patient-reported outcome measures (PROMs) are a gold standard for measuring therapeutic outcomes in research. Extending their use to inform clinical care decisions, determine the appropriateness of therapeutic choices, and assess healthcare quality is attractive but will require our professional community to establish valid estimates of minimal and substantial clinical improvements. QUESTIONS/PURPOSES: The purposes of this study were (1) to assess the validity of estimates for the minimal clinically important difference (MCID) calculated using distribution- and anchor-based methods by determining whether they exceed the minimal detectable change (MDC) for the Hip Disability and Osteoarthritis Outcome Score (HOOS) and the Knee Injury and Osteoarthritis Outcome Score (KOOS) domains, the HOOS, joint replacement (JR) and the KOOS, JR among patients who underwent THA or TKA; (2) to determine substantial clinical benefit thresholds for the HOOS and KOOS domains, the HOOS, JR, and the KOOS, JR among patients who underwent THA or TKA; and (3) to assess the proportions of patients who underwent THA or TKA who achieved an MCID for the HOOS and KOOS domains, HOOS, JR, and KOOS, JR based on distribution-based and anchor-based methods as well as the percentages of patients who achieved substantial clinical benefit using the anchor-based method. METHODS: Medicare patients enrolled in our institutional joint replacement registry who subsequently underwent THA (n = 2323) or TKA (n = 2630) between 2007 and 2012 completed HOOS or KOOS preoperatively and 2 years postoperatively. Short-form joint replacement (JR) versions of each PROM were derived from the full PROMs. Of all eligible patients, 78% (3161 of 4080) of THAs and 74% of TKAs (3815 of 5156) consented to join the registry and completed a baseline survey, 88% (2796 of 3161) of THAs and 85% (3230 of 3815) of TKAs were eligible for followup survey administration, and 83% of THAs (2323 of 2796) and 81% (2630 of 3230) of TKAs returned 2-year surveys. For each HOOS domain, KOOS domain, HOOS, JR, and KOOS, JR, we calculated the calibration variation of the instrument (MDC) with confidence intervals (CIs) reflecting 80% (MDC80), 90% (MDC90), and 95% (MDC95) certainty; we calculated the smallest difference joint health patients might detect (MCID) using distribution- and anchor-based approaches and the difference that can be considered a large improvement in joint health (substantial clinical benefit) using an anchor-based approach. RESULTS: Patients undergoing THA were 57% female with a mean (± SD) age of 73 ± 6 years, whereas patients undergoing TKA were 63% female with a mean age of 74 ± 6 years. Depending on the CI chosen for the MDC, values ranged from 7 to 16 for the HOOS and KOOS domains and the JRs. The MCIDs ranged from 6 to 9 for the distribution-based approach and 7 to 36 for the anchor-based approach. All HOOS and KOOS domains and all JR scores are scores from 0 (worst joint health) to 100 (best joint health). The MCIDs calculated using the distribution-based approach were not valid, because they were lower than the MDC for all HOOS/KOOS domains and both JRs at every confidence level. The anchor-based receiver operating characteristic approach, on the other hand, resulted in MCIDs exceeding MDC80 for seven of eight HOOS/KOOS domains and MDC95 for both JR scores. For all domains and JR versions, substantial clinical benefits ranged from 15 to 36, exceeding MDC95 in all domains and JR scores. Across HOOS and KOOS domains as well as the JR, the proportion of patients undergoing THA who achieved an MCID ranged from 77% to 95% with the distribution-based method and from 67% to 96% using the anchor-based method. The proportion achieving substantial clinical benefit ranged from 67% to 85%. CONCLUSIONS: The MDC and MCID differ greatly based on assumptions and methods used. The MCID anchor-based approach had superior construct and face validity compared with the MCID distribution-based approach, which never exceeded even small MDCs. Achieving consensus about standard definitions of meaningful improvement will be necessary to maximize utility of these PROMs to inform clinical care or performance measurement. LEVEL OF EVIDENCE: Level III, diagnostic study.
Subject(s)
Arthroplasty, Replacement, Hip/statistics & numerical data , Arthroplasty, Replacement, Knee/statistics & numerical data , Minimal Clinically Important Difference , Patient Reported Outcome Measures , Aged , Aged, 80 and over , Female , Humans , Male , Medicare , Registries , Reproducibility of Results , Treatment Outcome , United StatesSubject(s)
Internal Medicine , Patient Reported Outcome Measures , Humans , Quality Improvement , United StatesABSTRACT
BACKGROUND: Discharge destination, either home or skilled care facility, after total knee arthroplasty (TKA) may be associated with significant variation in postacute care outcomes. The purpose of this study was to characterize the 30-day postdischarge outcomes after primary TKA relative to discharge destination. METHODS: All primary unilateral TKAs performed for osteoarthritis from 2011-2014 were identified in the National Surgical Quality Improvement Program database. Propensity scores based on predischarge characteristics were used to adjust for selection bias in discharge destination. Propensity-adjusted multivariable logistic regressions were used to examine associations between discharge destination and postdischarge complications. RESULTS: Among 101,256 primary TKAs identified, 70,628 were discharged home and 30,628 to skilled care facilities. Patients discharged to facilities were more frequently were female, older, higher body mass index class, higher Charlson comorbidity index and American Society of Anesthesiologists scores, had predischarge complications, received general anesthesia, and classified as nonindependent preoperatively. Propensity adjustment accounted for this selection bias. Patients discharged to skilled care facilities after TKA had higher odds of any major complication (odds ratio = 1.25; 95% confidence interval, 1.13-1.37) and readmission (odds ratio = 1.81; 95% confidence interval, 1.50-2.18). Skilled care was associated with increased odds for respiratory, septic, thromboembolic, and urinary complications. Associations with death, cardiac, and wound complications were not significant. CONCLUSION: After controlling for predischarge characteristics, discharge to skilled care facilities vs home after primary TKA is associated with higher odds of numerous complications and unplanned readmission. These results support coordination of care pathways to facilitate home discharge after hospitalization for TKA whenever possible.
Subject(s)
Arthroplasty, Replacement, Knee/statistics & numerical data , Patient Discharge/statistics & numerical data , Postoperative Complications/epidemiology , Skilled Nursing Facilities/statistics & numerical data , Aged , Anesthesia, General , Databases, Factual , Female , Humans , Inpatients , Logistic Models , Male , Middle Aged , Odds Ratio , Propensity Score , Quality Improvement , United States/epidemiologyABSTRACT
BACKGROUND: Discharge disposition accounts for significant variability in costs after elective total hip arthroplasty (THA). Therefore, institutions must evaluate the short-term clinical outcomes associated with postdischarge care options. The present study intends to characterize the associations between short-term morbidity after primary THA and discharge destination. METHODS: Primary elective unilateral THA cases performed for osteoarthritis were identified in the American College of Surgeons National Surgical Quality Improvement Program registry from 2011 to 2014. Propensity scores were used to adjust for selection bias in discharge destination, based on demographics, obesity class, preoperative functional status, modified Charlson comorbidity index, American Society of Anesthesiologists (ASA) class, and the presence of predischarge complications. Propensity-adjusted multivariate logistic regressions were used to examine associations between discharge destination and postdischarge complications, controlling for selection bias based on observable patient characteristics. RESULTS: Among 54,837 THA cases included in the study, 40,576 (74%) were discharged home, and 14,261 (26%) were discharged to inpatient facilities. In multivariate propensity-adjusted analyses, patients discharged to continued inpatient care after THA were more likely to have septic complications (odds ratio, 2.34; 95% confidence interval, 1.58-3.45), urinary complications (1.51; 1.21-1.90), readmission (1.44; 1.29-1.59), wound complications (1.31; 1.09-1.57), and respiratory complications (1.93; 1.21-3.07). CONCLUSION: Discharge to continued inpatient care following THA is associated with increased odds of postdischarge morbidity and unplanned readmission, after propensity score adjustment for predischarge characteristics. Additional research is needed on the impact of devoting resources toward facilitating discharge to home after THA.
Subject(s)
Arthroplasty, Replacement, Hip/adverse effects , Patient Discharge/statistics & numerical data , Postoperative Complications/epidemiology , Residential Facilities/statistics & numerical data , Aged , Elective Surgical Procedures , Female , Humans , Inpatients , Logistic Models , Male , Middle Aged , Morbidity , Multivariate Analysis , Odds Ratio , Postoperative Complications/etiology , Propensity Score , Quality Improvement , Registries , Retrospective Studies , Risk Factors , United States/epidemiologyABSTRACT
BACKGROUND: Although consumers purchasing health plans in the new Health Insurance Marketplace will be provided information on the cost and quality of participating health plans, it is unclear whether the state-wide plan quality averages that will be reported will accurately represent quality at the pricing region level where care will be received. OBJECTIVES: To evaluate whether currently reported state-wide health plan quality scores accurately represent quality within pricing regions established for the Health Insurance Marketplace. RESEARCH DESIGN: Observational, historical cohort study using health plan administrative and pharmacy data. SUBJECTS: A total of 5.2 million members enrolled in the preferred provider organization health plans of 1 large commercial California insurer in 2012. MEASURES: State-wide and pricing region performance on each of the 17 Healthcare Effectiveness Data and Information Set (HEDIS) measures. RESULTS: Across the 17 measures assessed in each of the 19 pricing regions, scores were statistically different (P<0.05) than the overall plan rate for 176 (54%). Variations in scores across regions were observed for each measure ranging from 6.4-percentage points for engagement in treatment for people with dependence of alcohol or other drugs to 47.2-percentage points for appropriate testing for pharyngitis among children. CONCLUSIONS: Quality scores in California vary greatly across geographic regions. Statewide averages may misrepresent the quality of care that consumers are likely to receive within a geographic area making difficult assessments about the value of the health care.
Subject(s)
Health Insurance Exchanges , Preferred Provider Organizations/economics , Preferred Provider Organizations/standards , Quality Indicators, Health Care , State Health Plans/economics , State Health Plans/standards , California , Health Services Research , Humans , United StatesABSTRACT
ABSTRACT: Although well-accepted clinical practice guidelines exist for the diagnosis of prosthetic joint infection (PJI), little is known about the quality of diagnosis for PJI. The identification of quality gaps in the diagnosis of PJI would facilitate the development of care structures and processes to shorten time to diagnosis and reduce the significant morbidity, mortality, and economic burden associated with this condition. Hence, we sought to develop valid clinical quality measures to improve the timeliness and accuracy of PJI diagnosis. We convened a nine-member multidisciplinary national panel of PJI experts including orthopedic surgeons, infectious disease specialists, an emergency medicine physician, and a patient previously treated for PJI to review, discuss, and rate the validity of proposed measures using a modification of the RAND-UCLA appropriateness method. In total, 57 permutations of six proposed measures were rated. Populations considered to be at high enough risk for PJI that certain care processes should always be performed were identified by the panel. Among the proposed quality measures, the panel rated five as valid. These novel clinical quality measures could provide insight into care gaps in the diagnosis of PJI.
Subject(s)
Arthroplasty, Replacement , Prosthesis-Related Infections , Humans , Prosthesis-Related Infections/diagnosis , Evidence-Based MedicineABSTRACT
The intersection of big data and artificial intelligence (AI) has resulted in advances in numerous areas, including machine learning, computer vision, and natural language processing. Although there are many potentially transformative applications of AI in health care, including precision medicine, this industry has been slow to adopt these technologies. At the same time, the operations of health care have historically been system-directed and physician-directed rather than patient-centered. The application of AI to patient-reported outcome measures (PROMs), which provide insight into patient-centered health outcomes, could steer research and healthcare delivery toward decisions that optimize outcomes important to patients. Historically, PROMs have only been collected within research registries. However, the increasing availability of PROMs within electronic health records has led to their inclusion in big data ecosystems, where they can inform or be informed by other data elements. The use of big data to analyze PROMs can help establish norms, evaluate data distribution, and determine proportions of patients achieving change or threshold standards. This information can be used for benchmarking, risk adjustment, predictive modeling, and ultimately improving the health of individuals and populations.
Subject(s)
Artificial Intelligence , Ecosystem , Humans , Machine Learning , Big Data , Patient Reported Outcome MeasuresABSTRACT
BACKGROUND: Despite the increasing use of patient-reported outcome measures (PROMs), the methodology used to evaluate clinically significant postoperative outcomes after total knee arthroplasty (TKA) is variable. The review aimed to survey studies with identified PROM-based metrics of clinical efficacy and the assessment procedures after TKA. METHODS: The MEDLINE database was queried from 2008-2020. Inclusion criteria were: full texts, English language, primary TKA with minimum one-year follow-up, use of metrics for assessing clinical outcomes with PROMs, and primary derivations of metrics. The following PROM-based metrics were identified: minimal clinically important difference (MCID), minimum detectable change (MDC), patient acceptable symptom state (PASS), and substantial clinical benefit (SCB). Study design, PROM value data, and methods of derivation for metrics were recorded. RESULTS: We identified 18 studies (including 46,173 patients) that met the inclusion criteria. Across these studies, 10 different PROMs were employed, and MCID was derived in 15 studies (83%). The MCID was calculated using anchor-based techniques in nine studies (50%) and distribution techniques in eight studies (44%). PASS values were presented in two studies (11%) and SCB in one study (6%) using an anchor-based method; MDC was derived in four studies (22%) using the distribution method. CONCLUSION: There is variability in the TKA literature with respect to the definition and derivation of measurements of clinically significant outcomes. Standardization of these values may have implications for optimal case selection and PROM-based quality measurement, ultimately improving patient satisfaction and outcomes.