Population-Based Applications and Analytics Using Patient-Reported Outcome Measures.

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.

Patient-Reported Outcomes Measurement Information System Mobility Computerized Adaptive Testing Maintains High Correlation and Low Test Burden Compared With Legacy Hip-Specific Instruments in Patients Undergoing Hip Arthroscopy for Femoroacetabular Impingement.

PURPOSE: To evaluate the reliability, construct validity, and responsiveness of the lower extremity-specific Patient-Reported Outcomes Measurement Information System (PROMIS) Mobility (MO) bank in patients who underwent hip arthroscopic surgery for femoroacetabular impingement. METHODS: Patients who underwent primary hip arthroscopic surgery at a large academic musculoskeletal specialty center between November 2019 and November 2020 completed the following baseline and 6-month measures: PROMIS MO, PROMIS Pain Interference (PI), PROMIS Physical Function (PF), modified Harris Hip Score, International Hip Outcome Tool 33, visual analog scale, and Single Assessment Numeric Evaluation. Construct validity was evaluated using Spearman correlation coefficients. The number of questions until completion was recorded as a marker of test burden. The percentage of patients scoring at the extreme high (ceiling) or low (floor) for each measure was recorded to measure inclusivity. Responsiveness was tested by comparing differences between baseline and 6-month measures, controlling for age and sex, using generalized estimating equations. Magnitudes of responsiveness were assessed through the effect size (Cohen d). RESULTS: In this study, 660 patients (50% female patients) aged 32 ± 14 years were evaluated. PROMIS MO showed a strong correlation with PROMIS PF (r = 0.84, P < .001), the International Hip Outcome Tool 33 (r = 0.73, P < .001), PROMIS PI (r = -0.76, P < .001), and the modified Harris Hip Score (r = 0.73, P < .001). Neither PROMIS MO, PROMIS PI, nor PROMIS PF met the conventional criteria for floor or ceiling effects (≥15%). The mean number of questions answered (± standard deviation) was 4.7 ± 2.1 for PROMIS MO, 4.1 ± 0.6 for PROMIS PI, and 4.1 ± 0.6 for PROMIS PF. From baseline to 6 months, the PROMIS and legacy measures exhibited significant responsiveness (P < .05), with similar effect sizes between the patient-reported outcome measures. CONCLUSIONS: This longitudinal study reveals that in patients undergoing hip arthroscopy, PROMIS MO computerized adaptive testing maintains high correlation with legacy hip-specific instruments, significant responsiveness to change, and low test burden compared with legacy measures, with no ceiling or floor effects at 6-month postoperative follow-up. LEVEL OF EVIDENCE: Level IV, retrospective case series.

Satisfaction With the Process vs Outcome of Care in Total Hip and Knee Arthroplasty.

BACKGROUND: Patient satisfaction, with both process of care and outcome of care, is critical for measuring the quality and value of elective procedures such as arthroplasty. The purpose of this study is to evaluate the correlation between early postoperative satisfaction with the process of care and 2-year satisfaction with the outcome of care after total hip arthroplasty (THA) and total knee arthroplasty (TKA). METHODS: We retrospectively analyzed data from institutional arthroplasty registries. Satisfaction with the outcome of care was measured on a scale of 0-100 by a previously validated questionnaire administered 2 years postoperatively. Patient satisfaction with the process of care was measured by the Press Ganey (PG) inpatient survey, also scored 0-100. We examined the correlation between these 2 measures of satisfaction in patients who underwent primary THA or TKA. RESULTS: In total, 721 TKA patients and 760 THA patients underwent surgery and completed both the PG survey and 2-year satisfaction questionnaire. The mean age was 65.1 years with a mean body mass index of 28.8 and 56% were female. The mean PG survey score for the entire cohort was 95.6. The mean 2-year satisfaction score was 90.3. The Spearman correlation coefficient between the PG survey and the 2-year satisfaction survey was 0.23 for TKA patients (P < .001) and 0.13 for THA patients (P < .001). CONCLUSION: We found a weak correlation between the measurement of satisfaction with the process of care surrounding hip and knee arthroplasty using the PG survey and measurement of satisfaction with the outcome of care after arthroplasty using a validated 2-year satisfaction instrument.

Preoperative Evaluation of the Lower Extremity-Specific PROMIS Mobility Bank in Patients with ACL Tears.

PURPOSE: The purpose of this study was to preoperatively assess the Patient-Reported Outcomes Measurement Information System (PROMIS) Item Bank v2.0-Mobility in patients with anterior cruciate ligament (ACL) tears to (1) determine construct validity by correlating with legacy patient-reported outcomes (PROs), (2) evaluate test burden, (3) determine the presence of floor or ceiling effects, and (4) revisit the conventional threshold for inclusiveness (floor/ceiling effects) in the modern era of computer adaptive testing (CAT)-based PROs. METHODS: Patients at a large academic musculoskeletal specialty center diagnosed with ACL tears indicated for surgery were administered the following outcomes measures before surgery: PROMIS Mobility CAT, PROMIS Pain Interference CAT (PROMIS PI), International Knee Documentation Committee (IKDC), the Marx Knee Activity Rating Scale (Marx), and Single Assessment Numeric Evaluation (SANE). Construct validity was evaluated using Spearman correlation coefficients. Correlation strengths were defined as high (≥0.7), high-moderate (0.61-0.69), moderate (0.4-0.6), moderate-weak (0.31-0.39) and weak (≤0.3). Number of questions to completion were recorded as a marker of test burden. The percentage of patients scoring at the extreme high (ceiling) or low (floor) of each measure was recorded to measure inclusivity. RESULTS: A total of 1126 patients were evaluated. The mean number of questions answered (± standard deviation) was 4.7 ± 2.1 for PROMIS Mobility and 4.5 ± 1.9 for PROMIS PI. PROMIS Mobility demonstrated a high correlation with IKDC, (r = 0.81, P < .001), a high-moderate correlation with PROMIS PI (r = -0.63), and a moderate correlation with SANE (r = 0.46, P < .01). Neither PROMIS Mobility nor PROMIS PI met conventional criteria for floor or ceiling effects (>15%). CONCLUSIONS: The PROMIS Mobility measure maintains construct validity, because its scores correlate strongly with other PROs measuring physical function with high efficiency among preoperative patients with ACL injuries. Although ceiling effects of PROMIS Mobility CAT were below the conventional significance threshold of 15% at the preoperative timepoint in this population, this study provides critical feedback for redesigning the Mobility bank. LEVEL OF EVIDENCE: Level III (Diagnostic study).

Impact of COVID-19 on vulnerable patients with rheumatic disease: results of a worldwide survey.

OBJECTIVE: There is emerging evidence that COVID-19 disproportionately affects people from racial/ethnic minority and low socioeconomic status (SES) groups. Many physicians across the globe are changing practice patterns in response to the COVID-19 pandemic. We sought to examine the practice changes among rheumatologists and what they perceive the impact to be on their most vulnerable patients. METHODS: We administered an online survey to a convenience sample of rheumatologists worldwide during the initial height of the pandemic (between 8 April and 4 May 2020) via social media and group emails. We surveyed rheumatologists about their opinions regarding patients from low SES and racial/ethnic minority groups in the context of the COVID-19 pandemic. Mainly, what their specific concerns were, including the challenges of medication access; and about specific social factors (health literacy, poverty, food insecurity, access to telehealth video) that may be complicating the management of rheumatologic conditions during this time. RESULTS: 548 rheumatologists responded from 64 countries and shared concerns of food insecurity, low health literacy, poverty and factors that preclude social distancing such as working and dense housing conditions among their patients. Although 82% of rheumatologists had switched to telehealth video, 17% of respondents estimated that about a quarter of their patients did not have access to telehealth video, especially those from below the poverty line. The majority of respondents believed these vulnerable patients, from racial/ethnic minorities and from low SES groups, would do worse, in terms of morbidity and mortality, during the pandemic. CONCLUSION: In this sample of rheumatologists from 64 countries, there is a clear shift in practice to telehealth video consultations and widespread concern for socially and economically vulnerable patients with rheumatic disease.

Clinical Experience with COVID-19 at a Specialty Orthopedic Hospital Converted to a Pandemic Overflow Field Hospital.

BACKGROUND: COVID-19, the illness caused by the novel coronavirus, SARS-CoV-2, has sickened millions and killed hundreds of thousands as of June 2020. New York City was affected gravely. Our hospital, a specialty orthopedic hospital unaccustomed to large volumes of patients with life-threatening respiratory infections, underwent rapid adaptation to care for COVID-19 patients in response to emergency surge conditions at neighboring hospitals. PURPOSES: We sought to determine the attributes, pharmacologic and other treatments, and clinical course in the cohort of patients with COVID-19 who were admitted to our hospital at the height of the pandemic in April 2020 in New York City. METHODS: We conducted a retrospective observational cohort study of all patients admitted between April 1 and April 21, 2020, who had a diagnosis of COVID-19. Data were gathered from the electronic health record and by manual chart abstraction. RESULTS: Of the 148 patients admitted with COVID-19 (mean age, 62 years), ten patients died. There were no deaths among non-critically ill patients transferred from other hospitals, while 26% of those with critical illness died. A subset of COVID-19 patients was admitted for orthopedic and medical conditions other than COVID-19, and some of these patients required intensive care and ventilatory support. CONCLUSION: Professional and organizational flexibility during pandemic conditions allowed a specialty orthopedic hospital to provide excellent care in a global public health emergency.

Assessment of a Satisfaction Measure for Use After Primary Total Joint Arthroplasty.

BACKGROUND: Patient satisfaction after total hip (THA) and total knee arthroplasty (TKA) is a core outcome selected by the Outcomes Measurement in Rheumatology. Up to 20% of THA/TKA patients are dissatisfied. Improving patient satisfaction is hindered by the lack of a validated measurement tool that can accurately measure change. METHODS: The psychometric properties of a proposed satisfaction instrument, consisting of 4 questions rated on a Likert scale, scored 1-100, were tested for validity, reliability, and sensitivity to change using data collected between 2007 and 2011 in an arthroplasty registry. RESULTS: We demonstrated construct validity by confirming our hypothesis; satisfaction correlated with similar constructs. Satisfaction correlated moderately with pain relief (TKA ρ = 0.61, THA ρ = 0.47) and function (TKA ρ = 0.65, THA ρ = 0.51) at 2 years; there was no correlation with baseline/preoperative pain/function values, as expected. Overall Cronbach's alpha >0.88 confirmed internal consistency. Test-retest reliability with weighted kappa ranged 0.60-0.75 for TKA and 0.36-0.56 for THA. Hip disability and Osteoarthritis Outcome Score/Knee injury and Osteoarthritis Outcome Scores quality of life improvement (>30 points) corresponds to a mean satisfaction score of 93.2 (standard deviation, 11.5) after THA and 90.4 (standard deviation, 13.8) after TKA, and increasing relief of pain and functional improvement increased the strength of their association with satisfaction. The satisfaction measure has no copyright and is available free of cost and represents minimal responder burden. CONCLUSION: Patient satisfaction with THA/TKA can be measured with a validated 4-item questionnaire. This satisfaction measure can be included in a total joint arthroplasty core measurement set for total joint arthroplasty trials.

Validation of the Hip Disability and Osteoarthritis Outcome Score and Knee Injury and Osteoarthritis Outcome Score Pain and Function Subscales for Use in Total Hip Replacement and Total Knee Replacement Clinical Trials.

BACKGROUND: Total hip replacement (THR)/total knee replacement (TKR) studies do not uniformly measure patient centered domains, pain, and function. We aim to validate existing measures of pain and function within subscales of standard instruments to facilitate measurement. METHODS: We evaluated baseline and 2-year pain and function for THR and TKR using Hip disability and Osteoarthritis Outcome Score (HOOS)/Knee Injury and Osteoarthritis Outcome Score (KOOS), with primary unilateral TKR (4796) and THR (4801). Construct validity was assessed by correlating HOOS/KOOS pain and activities of daily living (ADL), function quality of life (QOL), and satisfaction using Spearman correlation coefficients. Patient relevant thresholds for change in pain and function were anchored to improvement in QOL; minimally clinically important difference (MCID) corresponded to "a little improvement" and a really important difference (RID) to a "moderate improvement." Pain and ADL function scores were compared by quartiles using Kruskal-Wallis. RESULTS: Two-year HOOS/KOOS pain and ADL function correlated with health-related QOL (KOOS pain and Short Form 12 Physical Component Scale ρ = 0.54; function ρ = 0.63). Comparing QOL by pain and function quartiles, the highest levels of pain relief and function were associated with the most improved QOL. MCID for pain was estimated at ≥20, and the RID ≥29; MCID for function ≥14, and the RID ≥23. The measures were responsive to change with large effect sizes (≥1.8). CONCLUSION: We confirm that HOOS/KOOS pain and ADL function subscales are valid measures of critical patient centered domains after THR/TKR, and achievable thresholds anchored to improved QOL. Cost-free availability and brevity makes them feasible, to be used in a core measurement set in total joint replacement trials.

Follow-Up of the Libby, Montana Screening Cohort: A 17-Year Mortality Study.

OBJECTIVE: To evaluate mortality patterns among participants in a community-based screening program for asbestos-related disease. METHODS: We calculated standardized mortality ratios (SMRs) and stratified results by exposure group (three occupational exposure groups, household contacts and residents without occupational asbestos exposure) and by radiographic abnormality presence. RESULTS: All-cause mortality (15.8%; 1,429/8,043) was statistically lower than expected. Asbestosis was statistically elevated in all exposure groups. Lung cancer was moderately associated with vermiculite miner/miller employment. Mesothelioma was elevated in that same exposure group and among residents. Systemic autoimmune disease mortality was also elevated. Radiographic parenchymal abnormalities were associated with lung cancer mortality. CONCLUSION: In addition to asbestos-related mortality in occupational exposure groups, this initial follow-up of this cohort also shows elevated mortality for some asbestos-related causes in non-occupational exposure groups.

9/11 Residential Exposures: The Impact of World Trade Center Dust on Respiratory Outcomes of Lower Manhattan Residents.

Thousands of lower Manhattan residents sustained damage to their homes following the collapse of the Twin Towers on 11 September 2001. Respiratory outcomes have been reported in this population. We sought to describe patterns of home damage and cleaning practices in lower Manhattan and their impacts on respiratory outcomes among World Trade Center Health Registry (WTCHR) respondents. Data were derived from WTCHR Wave 1 (W1) (9/2003⁻11/2004) and Wave 2 (W2) (11/2006⁻12/2007) surveys. Outcomes of interest were respiratory symptoms (shortness of breath (SoB), wheezing, persistent chronic cough, upper respiratory symptoms (URS)) first occurring or worsening after 9/11 W1 and still present at W2 and respiratory diseases (asthma and chronic obstructive pulmonary disease (COPD)) first diagnosed after 9/11 W1 and present at W2. We performed descriptive statistics, multivariate logistic regression and geospatial analyses, controlling for demographics and other exposure variables. A total of 6447 residents were included. Mean age on 9/11 was 45.1 years (±15.1 years), 42% were male, 45% had ever smoked cigarettes, and 44% reported some or intense dust cloud exposure on 9/11. The presence of debris was associated with chronic cough (adjusted OR (aOR) = 1.56, CI: 1.12⁻2.17), and upper respiratory symptoms (aOR = 1.56, CI: 1.24⁻1.95). A heavy coating of dust was associated with increased shortness of breath (aOR = 1.65, CI: 1.24⁻2.18), wheezing (aOR = 1.43, CI: 1.03⁻1.97), and chronic cough (aOR = 1.59, CI: 1.09⁻2.28). Dusting or sweeping without water was the cleaning behavior associated with the largest number of respiratory outcomes, such as shortness of breath, wheezing, and URS. Lower Manhattan residents who suffered home damage following the 9/11 attacks were more likely to report respiratory symptoms and diseases compared to those who did not report home damage.

Considerations Before Establishing an Environmental Health Registry.

Public health registries can provide valuable information when health consequences of environmental exposures are uncertain or will likely take long to develop. They can also aid research on diseases that may have environmental causes that are not completely well defined. We discuss factors to consider when deciding whether to create an environmental health registry. Those factors include public health significance, purpose and outcomes, duration and scope of data collection and availability of alternative data sources, timeliness, availability of funding and administrative capabilities, and whether the establishment of a registry can adequately address specific health concerns. We also discuss difficulties, limitations, and benefits of exposure and disease registries, based on the experience of the Agency for Toxic Substances and Disease Registry.

Surveillance for occupational respiratory diseases in developing countries.

The burden of chronic diseases, including occupational respiratory diseases (ORDs), is increasing worldwide. Nevertheless, epidemiological data on these conditions are scarce in most countries. Therefore, it is important to conduct surveillance to monitor ORDs, particularly in developing countries, where the working population is especially vulnerable and the health system infrastructure is usually weak. This article provides a general framework for the implementation of ORD surveillance in developing countries. The main objectives of surveillance are to describe incidence and prevalence of ORDs, as well as to identify sentinel events and new associations between occupational exposures and health outcomes. Diseases with high morbidity and mortality and those in which early diagnosis with standardized tests are available are especially suitable for surveillance activities. Simple strategies, preferably using existing resources and technology, are the best option for surveillance in developing countries. This article offers examples of specific surveillance systems that are in place in Brazil, China, Cuba, India, and South Africa.

The National ALS Registry: a recruitment tool for research.

INTRODUCTION: Subject recruitment is critical for understanding fatal diseases like ALS, however linking patients with researchers can be challenging. The U.S. population-based National ALS Registry allows recruitment of persons with ALS (PALS) for research opportunities. METHODS: The Registry's Research Notification Mechanism was used to recruit PALS aged ≥21 years; participants completed a Web-based epidemiologic survey. PALS (n = 2,232) were sent an email describing the study, and 268 surveys were completed. RESULTS: The mean age (± SD) of eligible participants was 57.7 ± 9.3 years for men and 61.5 ± 8.9 for women. Most were men (63%) and Caucasian (92%). Of 256 potentially eligible participants, 37.5% (n = 96) returned an authorization to disclose protected health information. ALS was confirmed for 94% (83/88) from physician responses. CONCLUSIONS: This analysis demonstrates the National ALS Registry's usefulness in recruiting PALS for research. This recruitment source can potentially foster the discovery of better treatment options and therapies, and of prevention strategies.

Impact of body mass index on the detection of radiographic localized pleural thickening.

RATIONALE AND OBJECTIVES: Subpleural fat can be difficult to distinguish from localized pleural thickening (LPT), a marker of asbestos exposure, on chest radiographs. The aims of this study were to examine the influence of body mass index (BMI) on the performance of radiograph readers when classifying LPT and to model the risk of false test results with varying BMI. MATERIALS AND METHODS: Subjects (n = 200) were patients being screened or treated for asbestos-related health outcomes. A film chest radiograph, a digital chest radiograph, and a high-resolution computed tomography (HRCT) chest scan were collected from each subject. All radiographs were independently read by seven B readers and scored using the International Labour Office system. HRCT scans, read by three experienced thoracic radiologists, served as the gold standard for the presence of LPT. We calculated measures of radiograph reader performance, including sensitivity and specificity, for each image modality. We also used logistic regression to estimate the probability of a false-positive and a false-negative result while controlling for covariates. RESULTS: The proportion of false-positive readings correlated with BMI. While controlling for covariates, regression modeling showed the probability of a false-positive result increased with increasing BMI category, younger age, not having pleural calcification, and among subjects not reporting occupational or household contact asbestos exposure. CONCLUSIONS: Clinicians should be cautious when evaluating radiographs of younger obese persons for the presence of asbestos-related pleural plaque, particularly in populations having an anticipated low or background prevalence of LPT.

Mesothelioma incidence in 50 states and the District of Columbia, United States, 2003-2008.

BACKGROUND: The decline in asbestos use in the United States may impact mesothelioma incidence. OBJECTIVE: This report provides national and state-specific estimates of mesothelioma incidence in the United States using cancer surveillance data for the entire US population. METHODS: Data from the National Program for Cancer Registries and the Surveillance, Epidemiology, and End Results program were used to calculate incidence rates and annual percent change. RESULTS: During 2003-2008, an average of 1.05 mesothelioma cases per 100 000 persons were diagnosed annually in the United States; the number of cases diagnosed each year remained level, whereas rates decreased among men and were stable among women. CONCLUSION: US population-based cancer registry data can be used to determine the burden of mesothelioma and track its decline. Even 30 years after peak asbestos use in the United States, 3200 mesothelioma cases are diagnosed annually, showing that the US population is still at risk.