Translate this page into:
Area Deprivation Index as a proxy for socioeconomic status in outpatient orthopaedic surgery patients – A prospective registry cross sectional study
, Daniel Polsky, Nathan N. O'Hara, Evan L. Honig, Sam Li, Craig Shul, Julio Jauregui, R. Frank Henn, Christopher G. Langhammer
⁎Corresponding author: Christopher G. Langhammer. clanghammer@som.umaryland.edu
-
Received: ,
Accepted: ,
This article was originally published by Reed Elsevier India Pvt. Ltd. and was migrated to Scientific Scholar after the change of Publisher.
Abstract
Abstract
We aimed to determine if Area Deprivation Index (ADI) is associated with self-reported metrics socioeconomic status (SES), and to assess the relationship between ADI and preoperative score on common patient reported outcome scores (PROS).
Patients presenting for outpatient orthopaedic surgery completed Patient-Reported Outcome Metric Information System (PROMIS) and joint-specific PROS. ADI was determined from geocoded home address. Sociodemographic data was collected from self-reported survey. Tests of association were used to describe the relationship between ADI and sociodemographic factors as well as the correlation between ADI and PROS. Extreme group analysis was used to examine which PROS may be subject to clinically meaningful variation.
ADI was associated with self-reported SES. ADI was correlated with score on all baseline PROS. Extreme group analysis showed that low SES was associated with clinically meaningful differences in some, but not all, PROS.
ADI is associated with self-reported measures of SES in an orthopaedic outpatient surgical population. Lower SES correlates with worse function to a clinically significant degree for some PROS. SES should be considered in the context of preoperative symptom severity in outpatient orthopaedic surgery patients. ADI may be a useful adjunct to self-reported measures of SES for this purpose.
1 Background
Social determinants of health include sociodemographic factors such as income, employment and education,1 which have assumed new prominence in health outcomes research because of their prognostic utility. Lower SES is associated with orthopaedic disease prevalence and symptom severity at multiple stages of treatment for orthopaedic conditions.2–7 Studies have demonstrated that the severity of baseline symptoms in orthopaedic clinic patients correlates with SES.7,8 Other studies have reported an association between ADI and postoperative function, satisfaction, and other PROS in postoperative orthopaedic patients. Some authors suggest a causal relationship between SES and outcome through a correlation with compliance with postoperative therapy protocols8–11 However, data about the utility of geography-based socioeconomic scores is heterogenous.12 While many studies identify an association between socioeconomic factors and outcomes, other work has demonstrated no effect.7 The role of SES in outcomes research remains controversial. The association of patient function and outcomes with SES appears to ultimately depend on multiple factors, including how SES is estimated, body region involved, and timing (initial presentation at clinic versus preoperative versus postoperative).13–16
Collecting SES patient data is costly. Self-reported data on employment, education, and income is the gold standard. This data is potentially sensitive and must be collected from the patient prospectively, making it impossible to use in retrospective studies. Recently, indirect indicators of socioeconomic status (SES) based on home address have become a popular method for estimating the net effect of multiple social factors on orthopaedic outcomes. These geography-based scores provide a quantitative composite indicator of a patients’ SES. The Area Deprivation Index (ADI) is one of the most widely used geography-based SES metrics. It incorporates 17 data points from the United States census that have been individually weighted to optimize prediction of all-cause mortality.8 Geography-based scores do not require the collection of sensitive patient data and are readily available for retrospective studies, overcoming many of the shortcomings of patient-reported SES data.17
To be a good candidate estimator of SES the measure should 1) capture information about known SES variables, and 2) correlate with outcomes. While prior studies have demonstrated association with symptom severity in an orthopaedic outpatient clinic population and postoperative patients, surgical indication and preoperative medical clearance imposes an intense SES-based selection bias occurring between presentation to clinic and presentation to surgery. Because of this selection bias, prior work on clinic and postoperative patients may not generalize to preoperative patients. This study aimed to address suitability of home address-based ADI as a measure of SES in patients undergoing outpatient orthopaedic surgery by determining if ADI is associated with self-reported metrics of SES and if ADI correlates with preoperative health status after the selection bias associated with surgical indication. We hypothesized that ADI will remain associated with self-reported measures of SES and higher ADI will remain associated with greater preoperative disability in patients throughout the selection process leading to outpatient orthopaedic surgery. This study is exploratory in nature. Because we do not try to explain a causal mechanism, we do not try to account for confounding in the relationship between ADI and PROs.
2 Methods
2.1 Data collection and management
All patients undergoing outpatient orthopaedic surgery at a single urban academic center from June 2015 to May 2021 were screened. Patients were excluded if they were younger than 12-years-old, did not speak English, did not have an email address, were institutionalized, or otherwise had a systematically incalculable ADI. In total, 3543 patients undergoing surgery were enrolled in a patient registry. Of those, 3207 patients had complete records available for analysis. This study received institutional review board approval.
Sociodemographic characteristics including income, employment, and education were self-reported. Electronic medical record review was used to retrieve relevant medical and operative history. The Charlson Comorbidity Index (CCI) was computed for quantifying comorbidity burden.18 Patients were stratified based on their operative site location, and a self-reported history of trauma before surgery.
2.2 Patient-reported outcome surveys
The computer-adaptive versions of PROMIS questionnaires for Physical Function, Pain Interference, Depression, Anxiety, Fatigue, and Social Satisfaction were administered to patients.19 One T score is reported for each domain, with a population mean of 50 and standard deviation of 10.19 Higher domain scores on Physical Function and Social Satisfaction imply better function, while high scores on Pain Interference, Depression, Anxiety, and Fatigue PROMIS domains imply worse function. Additional PROs collected preoperatively included the Numeric Pain Scale (NPS) for operative site and whole body pain, Tegner Activity Scale, and Marx Activity Rating Scales (MARS).19 Lastly, patients completed the most appropriate joint-specific survey based on their surgery: the American Shoulder and Elbow Surgeons (ASES) Score, International Knee Documentation Committee (IKDC) Score, or Brief Michigan Hand Questionnaire (BMHQ) Score.19
2.3 Area Deprivation Index
ADI is calculated as the standardized weighted average of 17 factors taken from United States census data at the level of the United States census block group. A census block group is a geographic area consisting of several city blocks with populations ranging from 600 to 3000 individuals and is the most granular division available using United States census data. The census factors included education level, white-collar occupation rate, median family income, income disparity, median home value, median rent, median monthly mortgage, home ownership rate, unemployment rate, family poverty rate, percentage of population 150 % below poverty rate, single parent household rate, household crowding, percentage of households with a motor vehicle, and percentage of households without a telephone. The weight for each factor was derived to fit all-cause mortality.17 The 17 weighted measures are summed, resulting in a base score for each census block group. The base score is then standardized on a national scale such that 1 corresponds with lowest deprivation (highest SES) and 100 corresponds with highest deprivation (lowest SES).20 Every individual's home address places them in a specific block group, associating them with that block group's ADI value.20,21 For this study, ADI was calculated using a geocoding algorithm scripted in Python to determine the location of each patient's home address based on street name and number and then identify within which census block group this address fell.
2.4 Statistical methods
Continuous variables were reported as means with standard deviation. Sociodemographic variables were categorical or were made categorical by binning continuous measures using values historically accepted as relevant to orthopaedics outcomes, and were reported as counts and percentages. Wilcoxon rank-sum tests and/or Kruskal-Wallis tests were used to assess the relationship between categorical sociodemographic variables and ADI. Post-hoc Wilcoxon analysis with Bonferroni corrections was performed for categorical variables with more than two subgroups. Normality of PROS was assessed with goodness-of-fit tests and a minority of outcome measurement scores were normally distributed, justifying nonparametric statistical tests. Spearman's correlation coefficient (Rs) was used to assess the strength of correlation between PROS and ADI. An extreme group analysis comparing mean PROS in the population deciles with the lowest and highest ADI using Wilcoxon rank sum test was performed to estimate the likely maximum PROS difference while excluding the central population where SES-effects are likely to be highly non-linear. Extreme group analysis was used over linear regression because of its very direct clinical interpretation as the difference that was observed between the tails of the observed population, rather than the difference between two hypothetical populations estimated based on regression coefficients. Multivariable analysis was not performed because of the exploratory nature of this investigation. All p-values were two-tailed and statistical significance was set at p < 0.05. Data was collected and housed in Research Electronic Data Capture (REDCap™, Boston, MA).22 Statistical analysis was performed using JMP®, Version 13 (SAS Institute Inc., Cary, NC).
3 Results
3.1 Sociodemographic factors, patient-specific factors, and operative factors association with ADI
Three thousand two hundred seven patients were included in final analysis. One thousand seven hundred forty-one (54.3 %) were male, with an average age of 45.2 years (SD: 17.5). One thousand nine hundred fifty-three (65.1 %) participants had a history of previous injury. The knee was the most common operative site (41.0 %) of participants. Other operative sites included shoulder (25.1 %), hand/wrist (20.9 %), elbow (6.1 %), hip (5.3 %), and foot/ankle (1.6 %). The National ADI mean and median were 33.7 and 28.0, respectively (range 1–100). Other sociodemographic and patient-specific factors of the population are listed in Table 1.
| N (%) | National ADI Score (1–100) | ||
| Mean (SD) | P-value | ||
| Demographic factors | |||
| Gender | |||
| Female | 1466 (45.7) | 36.9 (25.4) | <0.001 ∗ |
| Male | 1741 (54.3) | 31.0 (22.0) | |
| Race | |||
| Asian | 116 (3.7) | 20.3 (15.8) | <0.001 ∗ |
| Black | 1094 (35.1) | 47.9 (27.0) | |
| Other | 89 (2.9) | 35.0 (23.7) | |
| White | 1814 (58.3) | 25.8 (17.4) | |
| Ethnicity | |||
| Hispanic/Latino | 163 (5.3) | 29.3 (16.5) | 0.27 |
| NOT Hispanic/Latino | 2916 (94.7) | 34.1 (24.2) | |
| Marital status | |||
| Married | 1238 (41.0) | 29.0 (20.8) | <0.001 ∗ |
| Single | 1778 (59.0) | 36.8 (25.2) | |
| Age (years) | |||
| <65 | 2715 (84.7) | 33.9 (24.1) | 0.59 |
| ≥65 | 490 (15.3) | 32.6 (22.4) | |
| Socioeconomic factors | |||
| Education | |||
| College education | 1987 (66.0) | 29.8 (21.1) | <0.001 ∗ |
| High school graduate | 641 (21.3) | 45.4 (26.6) | |
| Less than high school | 384 (12.7) | 34.2 (26.0) | |
| Employment | |||
| Employed/retired | 1945 (64.8) | 32.0 (21.9) | <0.001 ∗ |
| Student | 526 (17.5) | 27.6 (20.8) | |
| Unable to work | 332 (11.1) | 48.4 (28.3) | |
| Unemployed | 199 (6.6) | 29.3 (29.1) | |
| Income | |||
| <$70,000 | 1345 (52.3) | 42.7 (25.8) | <0.001 ∗ |
| >$70,000 | 1227 (47.7) | 23.0 (16.1) | |
| Insurance type | |||
| Government | 816 (27.0) | 41.4 (26.5) | <0.001 ∗ |
| Private | 2202 (73.0) | 30.8 (22.0) | |
| Operative factors | |||
| Operative site location | |||
| Shoulder | 805 (25.1) | 34.7 (24.2) | <0.001 ∗ |
| Elbow | 197 (6.1) | 32.1 (23.4) | |
| Hand/wrist | 670 (20.9) | 38.3 (26.3) | |
| Hip | 169 (5.3) | 30.8 (19.3) | |
| Knee | 1316 (41.0) | 31.1 (22.3) | |
| Foot/ankle | 60 (1.6) | 41.4 (27.4) | |
| History of prior surgery at operative site | |||
| No | 2419 (76.5) | 34.2 (24.5) | 0.62 |
| Yes | 742 (23.5) | 32.5 (22.0) | |
| History of prior injury | |||
| No | 1048 (34.9) | 37.2 (25.3) | <0.001 ∗ |
| Yes | 1953 (65.1) | 31.8 (22.9) | |
| Patient-specific factors | |||
| Recreational drug use | |||
| No | 2799 (93.8) | 33.1 (23.5) | <0.001 ∗ |
| Yes | 186 (6.2) | 41.5 (27.4) | |
| Preoperative opioid use | |||
| No | 2437 (76.5) | 31.9 (22.8) | <0.001 ∗ |
| Yes | 748 (23.5) | 39.7 (26.2) | |
| Clinical history of depression or anxiety | |||
| No | 2681 (83.6) | 33.1 (23.7) | <0.001 ∗ |
| Yes | 526 (16.4) | 37.0 (24.3) | |
| Smoking | |||
| Daily | 393 (13.1) | 45.4 (27.6) | <0.001 ∗ |
| Never | 2030 (67.5) | 30.8 (22.1) | |
| Quit | 586 (19.5) | 35.8 (24.3) | |
| BMI (kg/m2) | |||
| Morbid obesity (BMI ≥40) | 229 (7.2)) | 45.7 (26.6) | <0.001 ∗ |
| Obese (30 ≤ BMI <40) | 1019 (31.8) | 37.2 (23.9) | |
| Healthy (BMI <30) | 1953 (61.0) | 30.5 (22.7) | |
| CCI (Mean = 1.44) | |||
| CCI >mean | 1297 (40.5) | 37.6 (25.3) | <0.001 ∗ |
| CCI <mean | 1908 (59.5) | 31.1 (22.4) | |
| ASA score | |||
| I | 1025 (33.0) | 28.3 (21.0) | <0.001 ∗ |
| II | 1776 (57.1) | 35.9 (24.5) | |
| III | 299 (9.6) | 39.1 (25.7) | |
| IV | 10 (0.3) | 48.2 (31.7) | |
Multiple demographic factors were associated with higher ADI, including African American race, single marital status, and female sex (all p < 0.001) (Table 1). Self-reported socioeconomic factors were associated with higher ADI including unemployment, lower income, and governmental insurance (all p < 0.001). Operative factors associated with higher ADI included operative site location and history of previous injury (all p < 0.001). Patient-specific factors associated with higher ADI included recreational drug use, preoperative opioid use, and worse baseline health status including higher CCI, body mass index, American Society of Anesthesiologists scores, and smoking status (all p < 0.001).
3.2 PROS association with ADI
Spearman's correlation coefficient showed that higher ADI correlated with worse scores on all baseline assessments (all p ≤ 0.005, Table 2). Special attention should be drawn to the PROMIS pain interference, fatigue, and social satisfaction instruments along with the NPS pain scales and the activity scales, all of which have Spearman's correlation coefficients >0.1, which is frequently used as a cutoff for non-negligible correlations. In the extreme group analysis, patients with the worst SES relative to those with the best SES had poorer baseline function across all domains except BMHQ (p ≤ 0.01) (Table 2). The largest mean differences across PROMIS domains were in Pain Interference (5.0), Fatigue (5.6), and Social Satisfaction (5.1). These domains were equal to or greater than the estimated PROMIS minimal clinically important difference (MCID) of ≥5.0. For activity rating scales, the largest differences were Tegner Activity Scale (10.6), MARS upper (18.2), and MARS lower (27.5). These scales were all larger than established MCIDs. The largest mean difference across joint-specific surveys was IKDC (10.7), larger than the established IKDC MCID of 10 (Table 2).
| Baseline PROS | Spearman's Correlation Coefficient with ADI | Extreme Group AnalysisMean PRO Score (SD) | Extreme GroupMean Difference (MCIDa) | P-value | ||||
| Mean (SD) | Median | (rs) | P-value | High SES Pop. Decile | Low SESPop. Decile | |||
| PROMIS Physical Function | 41.6 (9.3) | 41.7 | −0.09 | <0.001∗ | 42.5 (9.1) | 40.2 (9.0) | 2.3 (5) | <0.001∗ |
| PROMIS Pain Interference | 61.0 (7.5) | 61.5 | 0.19 | <0.001∗ | 58.8 (7.6) | 63.8 (7.0) | 5 (5) | <0.001∗ |
| PROMIS Fatigue | 52.2 (10.5) | 50.8 | 0.16 | <0.001∗ | 49.5 (10.3) | 55.1 (10.4) | 5.6 (5) | <0.001∗ |
| PROMIS Social Satisfaction | 42.0 (9.6) | 41.2 | −0.13 | <0.001∗ | 44.2 (9.7) | 39.1 (8.8) | 5.1 (5) | <0.001∗ |
| PROMIS Anxiety | 55.1 (9.3) | 55.7 | 0.08 | <0.001∗ | 53.8 (8.8) | 56.4 (9.5) | 2.6 (5) | <0.001∗ |
| PROMIS Depression | 49.0 (9.5) | 48.5 | 0.06 | 0.001∗ | 48.4 (8.9) | 50.2 (10.3) | 1.8 (5) | 0.017∗ |
| NPS op site | 50.6 (29.2) | 50 | 0.20 | <0.001∗ | 39.4 (26.9) | 63.1 (29.2) | 23.7 (21.7) | <0.001∗ |
| NPS whole body | 17.3 (24.6) | 0 | 0.12 | <0.001∗ | 10.1 (18.1) | 24.7 (30.1) | 14.6 (N/A) | <0.001∗ |
| MODEMS Expectations | 85.3 (18.7) | 91.7 | −0.08 | <0.001∗ | 87.8 (16.8) | 80.5 (21.5) | 7.3 (N/A) | <0.001∗ |
| Tegner Activity Scale | 22.2 (20.8) | 20 | −0.16 | <0.001∗ | 27.2 (21.6) | 16.6 (20.2) | 10.6 (10) | <0.001∗ |
| MARS Upper | 52.8 (31.0) | 55 | −0.16 | <0.001∗ | 61.0 (26.5) | 42.8 (32.8) | 18.2 (12.5) | <0.001∗ |
| MARS Lower | 48.2 (38.4) | 50 | −0.18 | <0.001∗ | 64.6 (34.5) | 37.1 (37.2) | 27.5 (12.5) | <0.001∗ |
| IKDC | 41.1 (17.5) | 40.2 | −0.13 | <0.001∗ | 45.6 (16.6) | 34.9 (17.0) | 10.7 (10) | <0.001∗ |
| ASES | 17.5 (13.5) | 16.7 | −0.16 | <0.001∗ | 21.5 (14.9) | 14.2 (11.7) | 7.3 (12–17) | 0.001∗ |
| BMHQ | 45.1 (20.4) | 43.8 | −0.12 | 0.005∗ | 45.7 (21.3) | 40.0 (21.2) | 5.7 (7) | 0.17 |
4 Discussion
There is increasing awareness of the role SES plays in predicting health outcomes.23 The most effective way to account for SES in orthopaedic research remains controversial.24 ADI is a geography-based composite metric of SES that can be calculated from a patient's home address, making it resilient to reporting bias and easily accessible for retrospective research. However, ADI's predictive utility in health outcomes was validated for a general population and its calculation was optimized around all-cause mortality, which might not capture less extreme or more subjective variance in orthopaedic outcomes.17
The population of patients seeking outpatient orthopaedic care, especially the subset eventually undergoing outpatient orthopaedic procedures, is subject to intense socioeconomic selection bias.7 Each stage of screening patients for elective surgery introduces bias.25 To present in an outpatient clinic, patients must have the resources to establish care and travel to an appointment. Patients with a higher ADI may delay seeking care or may not have means to access the healthcare system. Before indicating patients for surgery, surgeons evaluate if patients can provide safe postoperative care for themselves and if they believe that the patients’ symptoms are rooted in their anatomy and not caused or exacerbated by other patient-specific factors. Finally, electronic registries introduce technological barriers that prevent low SES individuals from participating. This multi-level selection bias is evident in our surgical population, which is skewed toward patients with higher SES (mean ADI of 33.6, median ADI of 28.0).
Our findings demonstrate that in a population of patients undergoing outpatient orthopaedic surgery, ADI remains correlated with self-reported socioeconomic factors predictive of orthopaedic outcomes, including level of education, employment, income, and insurance type. Patients with some college education had a mean ADI of 29.8, while patients with only a high school degree had a much higher ADI of 45.4. The higher SES observed for patients with higher educational attainment is consistent with the known increase in earning capacity associated with college graduation.
ADI additionally captures variance associated with multiple demographic factors that are thought to correlate with SES. Married patients had an average ADI of 29.0, while single patients had an ADI of 36.8, in agreement with findings that individuals with lower SES are less likely to marry.26 Self-identified African Americans had an average ADI of 47.9 compared to 29.3 for Hispanic race, 25.8 for White race and 20.3 for Asian race, consistent with the known socioeconomic disparities affecting the African American community. ADI is also associated with patient-specific factors such as recreational drug use, smoking status, preoperative opioid use, and health status, in agreement with previous research.27,28
Finally, this study also shows that increased ADI is associated with worse function at the time of surgery measured by PROs including multiple PROMIS domains, NPS scores, and Tegner and MARS activity scales. For the extreme group analysis, we used a 5-point MCID for PROMIS domains based on previous studies proposing a 3 to 5 point MCID.8,9 Comparisons between the highest and lowest SES deciles in this cohort showed the greatest differences in PROMIS Pain Interference, Fatigue, and Social Satisfaction with differences of 5.0, 5.6 and 5.1, respectively. We also found higher NPS scores in the lowest SES group relative to highest. These differences were all larger than established MCID values, suggesting that pain and pain effects correlate with socioeconomics, and that the effect is large enough to be observed as a clinically meaningful difference in function that cannot be explained by their orthopaedic pathology alone. Activity was similarly associated with SES. Tegner Activity Scale and MARS upper and lower extremity surveys showed reduced activity levels that were above estimated MCID (estimated at 10.0 and 12.5, respectively) in the population with the highest ADI.29,30
Previous studies have demonstrated an association between ADI and symptom severity in orthopaedic outpatient clinics.8,9,31 Although ADI has been validated in a clinic population, no previous research has examined its correlation with socioeconomic, demographic, and patient-specific factors in the subset of patients selected from the larger pool of clinic patients to undergo outpatient surgery. This study expands previous examinations of the effects of SES on function to a new clinical environment: outpatient orthopaedic surgeries. Ultimately, ADI can be a convenient and effective metric for quantifying SES as it pertains to orthopaedic outcomes in patients presenting for outpatient surgery because it meets the following two criteria: 1) ADI captures aspects of self-reported sociodemographics and patient health status known to vary with SES, and 2) commonly used PROS are responsive to a degree reaching clinical significance to the range of clinically encountered ADIs.
An important strength of this study is the racial diversity of the cohort, consisting of 35.1 % African American participants. This number is high relative to many studies and is important to consider as African Americans have historically been underrepresented in research and are disproportionately burdened with low SES relative to other demographics.8,32 Another strength is the reverse-geocoding method used to calculate ADI, accurate to the level of street number. This provides more accurate estimations of a patient's SES than zip code-based methods. Using zip code-based metrics creates a volume averaging effect so severe it can reverse the direction of findings in some health outcomes studies.33
The study has notable limitations. It is exploratory in nature. Statistical tests were selected to examine suitability of ADI as a proxy for SES in orthopaedic research, especially considering the diversity of PROS that are now commonly used. The purpose of this study was not to identify components of ADI that can be targeted for intervention or are specific risk factors for low orthopaedic function, so we have deliberately limited discussion of causal mechanism. The findings should be considered within this context. While correlation between ADI and many PROs reaches a level of statistical significance with p < 0.001, the extent of correlation for each scale was Rs < 0.3. Although considered “low” in the context of the physical sciences, the social sciences and recent SES research in orthopaedics use a threshold for clinical relevance of Rs > 0.1.23 Also, a mean difference assessment in an extreme group analysis is prone to overstating the magnitude of effect. This effect magnification is mitigated by using population deciles (i.e., we use the top 10 % of the observed population vs. the bottom 10 % of the observed population) rather than SES deciles (i.e., ADI 0–10 vs. ADI 90–100). This was selected as a way of minimizing the effects of expected nonlinearities associated with mid-range SES and to provide an estimate of the observed maximum effect. It avoids providing an inflated estimate of a maximum effect over a range of SES not actually observed in clinical practice. We felt that using regression coefficients to estimate the effect of SES on symptom severity in the lowest 10 % of the SES scale is clinically uninterpretable, as this portion of the population is rarely indicated for outpatient orthopaedic surgery, or even seen in an outpatient office. As surgical providers we strive to provide excellent results, and naturally compare individual patient outcomes to our best outcomes. The findings from the selected analysis are readily interpreted within the context of daily clinical experience: 1 in 10 patients indicated for surgery may have a clinically obvious difference in symptom severity rooted in socioeconomics rather than orthopaedic pathology. An understanding of this observation may help providers find greater satisfaction in their daily patient interactions.
Reliable assessment of SES is important for contextualizing outcomes, improving patient-tailored preoperative counseling, comparing surgeon and center performance, and potentially modifying reimbursement rates according to patient sociodemographic factors. It is important to note the ethical implications of considering SES in outcomes research. SES undeniably effects our patients’ reported outcomes. Geography-based metrics like ADI may provide an objective way of accounting for SES in research studies. Conversely, if it is inappropriately applied, a metric like this may fuel additional discrimination against a demographic already facing disproportionate barriers to care. It is our moral imperative to understand these interactions in order to avoid introducing bias in decision making.
It is our hope that this paper helps define the scope of involvement SES may have on outcomes, so that future research can focus on developing targeted interventions to improve outcomes in vulnerable low SES populations. It is notable that some, but not all, PROS appear susceptible to SES-based moderation. Future research into which PROS are most effected by SES, as well as how they are affected, are required to understand this in greater detail. We feel this question may be most appropriately explored in future studies using a repeated measures design reporting postoperative score change for a panel of appropriately selected PROs.
5 Conclusion
ADI remains associated with known measures of self-reported SES in patients presenting for outpatient orthopaedic surgery despite selection bias intrinsic in this patient group. Higher ADI (lower SES) has a clinically meaningful association with preoperative function as measured by common PROS. Neighborhood-based metrics such as ADI may be used as a proxy measure for SES in orthopaedic research when collection of self-reported measures of SES is impractical, and should be considered an important patient-related factor in outcomes-based research.
Ethics approval
This study was approved by the local institutional review board.
Funding
This work was supported by the James Lawrence Kernan Hospital Endowment Fund, Incorporated: Grant # BL1941007WS.
Guardian/patient's consent statement
All patients provided informed consent to participate in the registry.
CRediT authorship contribution statement
Samir Kaveeshwar: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. Sania Hasan: Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Daniel Polsky: Data curation, Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Nathan N. O'Hara: Data curation, Methodology, Formal analysis, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Evan L. Honig: Formal analysis, Investigation, Methodology, Writing – review & editing. Sam Li: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Craig Shul: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. Julio Jauregui: Formal analysis, Investigation, Methodology, Writing – original draft, Writing – review & editing. R. Frank Henn: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing. Christopher G. Langhammer: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing.
References
- Socioeconomic status and health: how education, income, and occupation contribute to risk factors for cardiovascular disease. Am J Publ Health. 1992;82(6):816-820.
- [Google Scholar]
- Social determinants of health and patient-reported outcomes following total hip and knee arthroplasty in veterans. J Arthroplasty. 2020;35(9):2357-2362.
- [Google Scholar]
- Disparities in total hip arthroplasty outcomes: census tract data show interactions between race and community deprivation. J Am Acad Orthop Surg. 2018;26(21):e457-e464.
- [Google Scholar]
- Healthcare disparities among orthopedic trauma patients in the USA: socio-demographic factors influence the management of calcaneus fractures. J Orthop Surg Res. 2019;14:359.
- [Google Scholar]
- Do patient sociodemographic factors impact the PROMIS scores meeting the patient-acceptable symptom state at the initial point of care in orthopaedic foot and ankle patients? Clin Orthop Relat Res. 2019;477(11):2555-2565.
- [Google Scholar]
- Estimating the burden of musculoskeletal disorders in the community: the comparative prevalence of symptoms at different anatomical sites, and the relation to social deprivation. Ann Rheum Dis. 1998;57(11):649-655.
- [Google Scholar]
- The impact of socioeconomic status on outcomes in orthopaedic surgery. J Bone Joint Surg Am. 2020;102(5):428-444.
- [Google Scholar]
- What is the impact of social deprivation on physical and mental health in orthopaedic patients? Clin Orthop Relat Res. 2019;477(8):1825-1835.
- [Google Scholar]
- Is social deprivation associated with PROMIS outcomes after upper extremity fractures in children? Clin Orthop Relat Res. 2021;479(4):826-834.
- [Google Scholar]
- The impact of social deprivation on orthopaedic outpatient satisfaction using the press ganey outpatient medical practice survey. J Am Acad Orthop Surg. 2020;28(24):e1111-e1120.
- [Google Scholar]
- Socioeconomic deprivation and its adverse association with adolescent fracture care compliance. JB JS Open Access. 2020;5(2)
- [Google Scholar]
- Reporting of social deprivation in musculoskeletal trials: an analysis of 402 randomised controlled trials. Muscoskel Care. 2020 Oct;14
- [Google Scholar]
- Outcomes following large joint arthroplasty: does socio-economic status matter? BMC Muscoskel Disord. 2014;15:148.
- [Google Scholar]
- Socioeconomic status affects the Oxford knee score and short-form 12 score following total knee replacement. Bone Joint J. 2013;95-B:52-58.
- [Google Scholar]
- The influence of area level social deprivation on preoperative disease severity and postoperative outcomes following unicompartmental knee joint replacement. Knee. 2015;22(6):653-658.
- [Google Scholar]
- Social deprivation influences the epidemiology and outcome of proximal humeral fractures in adults for a defined urban population of Scotland. Eur J Orthop Surg Traumatol. 2014;24(7):1039-1046.
- [Google Scholar]
- Area deprivation and widening inequalities in US mortality, 1969-1998. Am J Publ Health. 2003;93(7):1137-1143.
- [Google Scholar]
- A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis. 1987;40(5):373-383.
- [Google Scholar]
- The Maryland Orthopaedic Registry (MOR): design and baseline characteristics of a prospective registry. J Clin Orthop Trauma. 2017;8(4):301-307.
- [Google Scholar]
- Introduction of an Area Deprivation Index measuring patient socioeconomic status in an integrated health system: implications for population health. EGEMS (Wash DC). 2016;4(3):1238.
- [Google Scholar]
- Neighborhood socioeconomic disadvantage is associated with multimorbidity in a geographically-defined community. BMC Publ Health. 2020;20(1):13.
- [Google Scholar]
- Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inf. 2009;42(2):377-381.
- [Google Scholar]
- How should we measure social deprivation in orthopaedic patients? Clin Orthop Relat Res. 2022;480(2):325-339.
- [Google Scholar]
- Editorial: beware of studies claiming that social factors are "independently associated" with biological complications of surgery. Clin Orthop Relat Res. 2019;477:1967-1969.
- [Google Scholar]
- Current concepts in orthopaedic care disparities. J Am Acad Orthop Surg. 2018;26(23):823-832.
- [Google Scholar]
- Socioeconomic status and intimate relationships. Annu Rev Psychol. 2021;72:391-414.
- [Google Scholar]
- Lower socioeconomic status is associated with worse preoperative function, pain, and increased opioid use in patients with primary glenohumeral osteoarthritis. J Am Acad Orthop Surg. 2020;28(7):287-292.
- [Google Scholar]
- Neighborhood socioeconomic status and substance use by U.S. adults. Drug Alcohol Depend. 2013;133(1):212-221.
- [Google Scholar]
- Clinical and functional outcomes of anterior cruciate ligament reconstruction with autologous hamstring tendon in patients aged 50 years or older. Arthroscopy. 2020;36(2):558-562.
- [Google Scholar]
- Change in KOOS and WOMAC scores in a young athletic population with and without anterior cruciate ligament injury. Am J Sports Med. 2018;46(7):1606-1616.
- [Google Scholar]
- Mental and physical health disparities in patients with carpal tunnel syndrome living with high levels of social deprivation. J Hand Surg Am. 2019;44(4):335.e331-335.e339.
- [Google Scholar]
- African Americans’ participation in clinical research: importance, barriers, and solutions. Am J Surg. 2007;193(1):32-39; discussion 40.
- [Google Scholar]