The complex interrelationship among physical health, mental health, and social health has gained the attention of the medical community in recent years. Poor social health, also called social deprivation, has been linked to more disease and a more-negative impact from disease across a wide variety of health conditions. However, it remains unknown how social deprivation is related to physical and mental health in patients presenting for orthopaedic care.
(1) Do patients living in zip codes with higher social deprivation report lower levels of physical function and higher levels of pain interference, depression, and anxiety as measured by Patient-Reported Outcomes Measurement Information System (PROMIS) at initial presentation to an orthopaedic provider than those from less deprived areas; and if so, is this relationship independent of other potentially confounding factors such as age, sex, and race? (2) Does the relationship between the level of social deprivation of a patient’s community and that patient’s physical function, pain interference, depression, and anxiety, as measured by PROMIS remain consistent across all orthopaedic subspecialties? (3) Are there differences in the proportion of individuals from areas of high and low levels of social deprivation seen by the various orthopaedic subspecialties at one large, tertiary orthopaedic referral center?
This cross-sectional evaluation analyzed 7500 new adult patients presenting to an orthopaedic center between August 1, 2016 and December 15, 2016. Patients completed PROMIS Physical Function-v1.2, Pain Interference-v1.1, Depression-v1.0, and Anxiety-v1.0 Computer Adaptive Tests. The Area Deprivation Index, a composite measure of community-level social deprivation, based on multiple census metrics such as income, education level, and housing type for a given nine-digit zip code was used to estimate individual social deprivation. Statistical analysis determined the effect of disparate area deprivation (based on most- and least-deprived national quartiles) for the entire sample as well as for patients categorized by the orthopaedic subspecialty providing care. Comparisons of PROMIS scores among these groups were based on an MCID of 5 points for each PROMIS domain (Effect size 0.5).
Patients living in zip codes with the highest levels of social deprivation had worse mean scores across all four PROMIS domains when compared with those living in the least-deprived quartile (physical function 38 +/- 9 versus 43 +/- 9, mean difference 4, 95% CI, 3.7–5.0; p < 0.001; pain interference 64 +/- 8 versus 60+/-8, mean difference -4, 95% CI, -4.8 to -3.7; p < 0.001; depression 50+/-11 versus 45+/-8, mean difference -5, 95% CI, -6.0 to -4.5; p < 0.001; anxiety 56+/-11 versus 50 +/-10, mean difference -6, 95% CI, -6.9 to -5.4; p < 0.001). There were no differences in physical function, pain interference, depression, or anxiety PROMIS scores between patients from the most- and least-deprived quartiles who presented to the subspecialties of spine (physical function, mean 35+/-7 versus 35+/-7; p = 0.872; pain interference, 67+/-7 versus 66+/-7; p = 0.562; depression, 54+/-12 versus 51 +/-10; p = 0.085; and anxiety, 60+/-11 versus 58 +/-9; p = 0.163), oncology (physical function, mean 33+/-9 versus 38 +/-13; p = 0.105; pain interference, 68+/-9 versus 64+/-10; p = 0.144; depression, 51+/-10 versus 52+/-13; p = 0.832; anxiety, 59+/-11 versus 59+/-10 p = 0.947); and trauma (physical function, 35+/-11 versus 32+/-10; p = 0.268; pain interference, 66+/-7 versus 67+/-6; p = 0.566; depression, 52+/-12 versus 53+/-11; p = 0.637; and anxiety, 59+/-12 versus 60+/-9 versus; p = 0.800). The social deprivation-based differences in all PROMIS domains remained for the subspecialties of foot/ankle, where mean differences ranged from 3 to 6 points on the PROMIS domains (p < 0.001 for all four domains), joint reconstruction where mean differences ranged from 4 to 7 points on the PROMIS domains (p < 0.001 for all four domains), sports medicine where mean differences in PROMIS scores ranged from 3 to 5 between quartiles (p < 0.001 for all four domains), and finally upper extremity where mean differences in PROMIS scores between the most- and least-deprived quartiles were five points for each PROMIS domain (p < 0.001 for all four domains). The proportion of individuals from the most- and least-deprived quartiles was distinct when looking across all seven subspecialty categories; only 11% of patients presenting to sports medicine providers and 17% of patients presenting to upper extremity providers were from the most-deprived quartile, while 39% of trauma patients were from the most-deprived quartile (p < 0.001).
Orthopaedic patients must be considered within the context of their social environment because it influences patient-reported physical and mental health as well as has potential implications for treatment and prognosis. Social deprivation may need to be considered when using patient-reported outcomes to judge the value of care delivered between practices or across specialties. Further studies should examine potential interventions to improve the perceived health of patients residing in communities with greater social deprivation and to determine how social health influences ultimate orthopaedic treatment outcomes.
In recent years, an expanding body of evidence has emphasized the inter-relationship between psychologic distress and poor orthopaedic outcomes [7, 18, 20, 25, 31, 32, 36, 37, 53]. However, more broadly across medicine, research suggests that not only are mental and physical health interrelated, but also that social health represents another facet of life, one often outside of the individual’s control, that may influence all other health domains . Increased social deprivation is associated with a greater prevalence of colorectal cancer, cardiac disease, weight gain, musculoskeletal pain as well as increased mortality rates [26, 30, 42, 48, 52]. Additionally, patients living in areas of increased social deprivation report increased levels of depressive symptoms as well as overall lower levels of self-reported well-being [16, 27, 33, 38, 43, 45].
Within the last decade, use of the Patient-Reported Outcomes Measurement Information System (PROMIS) has become increasingly more common as a method of assessing mental and physical health domains in patients, both singly and in groups. PROMIS physical function scores correlate well with traditional orthopaedic patient-reported outcomes, such as the QuickDASH for upper extremity patients and the short Musculoskeletal Assessment form for trauma patients. Various domains of PROMIS have demonstrated effectiveness in orthopaedic patient populations [1, 9, 12, 22, 41, 49, 51], providing an efficient way for orthopaedic physicians to assess patient perceptions regarding their physical and mental health and well-being. PROMIS assessments are not disease- or diagnosis-specific and thus represent a potential consensus measurement system to quantify the relative value of care delivered across medical and surgical specialties.
Although we have made strides in assessing patients’ psychologic well-being, the effect of social deprivation on physical and mental health in orthopaedic patients has not been well defined. A large proportion of orthopaedic care can be viewed as elective, in the sense that it is neither life-saving nor life-prolonging. In light of the potential for serious harm to patients in this setting, it seems appropriate to identify possible barriers to achieving good outcomes, and address them before definitive management. It is therefore critical to understand if and how social health affects patient-reported physical and mental health ratings and where interventions may be needed. Therefore, this study aimed to answer the following questions.
(1) Do patients living in zip codes with higher social deprivation report lower levels of physical function and higher levels of pain interference, depression, and anxiety as measured by PROMIS at initial presentation to an orthopaedic provider than those from less deprived areas; and if so, is this relationship independent of other potentially confounding factors such as age, gender, and race? (2) Does the relationship between the level of social deprivation of a patient’s community and that patient’s physical function, pain interference, depression, and anxiety, as measured by PROMIS remain consistent across all orthopaedic subspecialties? (3) Are there differences in the proportion of individuals from areas of high and low levels of social deprivation seen by the various orthopaedic subspecialties at one large, tertiary orthopaedic referral center?
Approval for this study was obtained from our institutional review board with a waiver of informed consent for all patients. This cross-sectional study collected data from 9737 new patient visits to the outpatient offices of a single, tertiary orthopaedic practice. Patients presented to five different clinical sites, two in zip codes with high levels of social deprivation, and three in zip codes with low levels of deprivation. Most patients were seen at two of the sites, one in a zip code of high deprivation, and one in a zip code of low deprivation. All specialty divisions are represented by at least one clinical site in an area of high deprivation, and one in an area of low deprivation, with the exception of trauma, where all clinical sites are in areas of high deprivation. Patients presented to any of the 61 providers in the department, as well as to the resident-run clinic site. Data were collected at the time of the visit for all new adult patients presenting between August 1, 2016 and December 15, 2016.
All adult (18 years of age or older) new patient visits to the practice during this study period were eligible for inclusion. During this study period, the PROMIS completion rate was 91% for all orthopaedic patients with completion of PROMIS Pain Interference and Physical Function required for study inclusion. Patients without a nine-digit zip code matched to their self-reported address or those whose address did not correlate with an Area Deprivation Index (ADI) number such as a PO Box were excluded. For patients presenting as a new patient to multiple providers within our group during the study dates, only the first visit was included with all subsequent visits removed from the data set. After applying all study participation requirements, 7500 patients remained eligible for inclusion (Fig. ).
As part of our institution’s standard practice, all patients completed the PROMIS Physical Function-v1.2, PROMIS Pain Interference-v1.1, PROMIS Depression-v1.0, and PROMIS Anxiety-v1.0 at presentation to their provider. Patients also completed a variety of demographic forms that include self-reported age, gender, race, and home address.
Our institution’s electronic medical record and administrative databases were queried for age, sex, race, home address (including nine-digit zip code), provider visited, and PROMIS Computer Adaptive Test (CAT) scores (pain interference, physical function, depression, and anxiety). Gender was obtained from the medical record and is self-reported as man or woman at the time of registration. Race is similarly self-reported at the time of registration. Patients were categorized as white, black, Asian, or other for study analysis. Home address is self-reported with the nine-digit zip code generated automatically during registration. Using the provider seen, participants were grouped by the orthopaedic subspecialty that provided care (foot/ankle, joint reconstruction [hip and knee], sports medicine, spine, trauma, oncology, upper extremity, and nonoperative providers).
The ADI quantified the social deprivation of each patient’s community according to the nine-digit home zip code. The ADI is a validated measure of social deprivation developed by George Singh that accounts for 17 US Census measures across the domains of poverty, education, housing, and employment, which mirror those used in other countries to measure social deprivation [8, 26, 46, 47]. Assessment of the index validity and reliability are detailed by Singh  in his paper describing creation of the ADI. The ADI of each geographic area is determined by the sum of each included socioeconomic factor multiplied by its predetermined factor coefficient; scores are then standardized so that the index mean is 100 with a SD of 20. After standardization, index values range from -53 to 129 with increasing values indicating increasing levels of social deprivation; a patient whose zip code translates to an ADI of 110 is from a more deprived area than someone with an ADI of 20. Socioeconomic disadvantage is a complex entity that indicates a combination of low income, limited education, poor living conditions, and decreased social support. Area deprivation affects both the individual and those around them . Using a composite index such as the ADI, which encompasses multiple domains that contribute to disadvantage, is therefore more robust than choosing a single metric such as income or education to assess social disadvantage . In addition to using each person’s address-based ADI score for analysis, we categorized patients according to established break points defining national quartiles of deprivation to allow for direct comparison between the most- and least-deprived groups of individuals, which is consistent with previous investigations of social deprivation [2, 10, 15, 24, 26, 47, 48]. While socioeconomic position and social status can be studied using a variety of metrics such as individual income, individual education level, or even composite metrics of these individual data points, we have chosen to use a zip-code based metric for deprivation as it serves as a proxy for individual social position while also characterizing the social condition of the surrounding community, which may have a further impact on health .
All patients presenting to our institution’s orthopaedic clinics are provided a computer tablet (iPad mini; Apple, Cupertino, CA, USA) at check-in that is preloaded with PROMIS Physical Function-v1.2, PROMIS Pain Interference-v1.1, PROMIS Depression-v1.0, and PROMIS Anxiety-v1.0. PROMIS scores are automatically generated and uploaded into the patient’s electronic health record on module completion. PROMIS CATs are scored such that a score of 50 represents the normative population mean with a SD of 10 . A higher score represents a higher level of a given metric such that a higher physical function score indicates better function, whereas a higher depression score indicates more depression. PROMIS has been well validated in many different patient populations, including a variety of orthopaedic populations [3, 21, 39, 49, 51]. Additionally, the PROMIS CATs, particularly physical function and pain interference, have been compared with multiple orthopaedic legacy measures of pain and function and found to correlate highly, often with less unexplained variance and less floor and ceiling effects than the traditional measures [12, 22, 51]. Pain interference is an assessment that measures how pain impacts an individual’s activities and function as opposed to simply measuring pain magnitude. Thus, pain interference blends elements of pain tolerance, coping, and functional demands.
Patient age was widely distributed in our study population (mean, 52 years; range, 18-98 years) and home addresses indicated residence in 33 different states (Table ). Most patients were white (85%). Area deprivation scores ranged from -53 to 128, a range comparable to the national range (-53 to 129). The median ADI was 99 (30th percentile nationally) indicating more patients from more socially advantaged areas. A total of 2799 patients (37%) were categorized within the least-deprived national quartile, and 1133 patients (15%) were categorized within the most-deprived national quartile.
Univariate descriptive statistics were produced to characterize continuous and categorical variables of interest: patient demographic data, PROMIS scores, ADI frequencies, and the percentage of visits to each subspecialty from the four ADI quartiles. Student’s t-tests compared mean PROMIS scores between individuals in the most- and least-deprived quartiles of the population, similar to other studies of social deprivation in which primary analysis compared the most- and least-deprived groups [26, 30, 42, 47, 48]. An estimated minimal clinically important difference (MCID) of 5 points on PROMIS assessments was used as a conservative proxy for clinically relevant between-group differences based on previous studies that have attempted to define MCID values for PROMIS short forms and CATs in various populations [1, 29, 50, 57]. The same statistical tests were performed to satisfy our secondary aim of comparing the impact of ADI among orthopaedic surgical subspecialties, excluding nonoperative providers. The Kruskal-Wallis test was used to compare the proportions of individuals in the most- and least-deprived quartiles presenting to the different subspecialties with pairwise comparison post hoc analysis using the Dunn-Bonferroni method. One-way analysis of variance was used to compare mean PROMIS scores among the different subspecialty groups with Bonferroni pairwise post-hoc analysis. Linear regression was used to model the impact of age, sex, race, subspecialty, and deprivation quartile on each PROMIS domain. All patients were included in the regression from all four quartiles. The regression models were based on forward entry of each of the independent variables with the α level set at 0.05 for entry into the model. Statistical analyses were performed using SPSS statistical software (Version 23; IBM, New York, NY, USA). Statistical significance was set at p < 0.05. Any missing data elements were excluded from analysis with no imputation performed.