Neighborhood Socioeconomic Disadvantage is Associated with Worse Urinary Risk Factors and Access to Medical Management for Urolithiasis

Introduction

Kidney stone disease is a chronic medical and surgical condition, with prevalence rates estimated to range from 7% to 13% in North America. ¹ Nearly one-half of patients with renal colic will experience a relapse at 10 years. ² Current guidelines recommend that high-risk or recurrent stone formers should be offered metabolic evaluation to help inform and monitor treatment protocols. ³ A 24-hour urine (24hU) testing helps direct dietary recommendations as well as initiate medications and counseling to prevent recurrent urolithiasis. ^4,5

In the United States, health disparities affect the incidence of chronic diseases and the ability of patients to seek and receive treatment. Patients with poor socioeconomic status (SES) are disproportionately affected by nephrolithiasis, and SES has been associated with larger stone burden, metabolic urine abnormalities, and recurrent disease. ^{6 –8} The neighborhood context and the environment in which a patient lives directly affect access to food, safety, education, health-related behaviors, and stress levels, and how disadvantaged a neighborhood is may influence health independently of a person’s SES. ⁹ Identifying these patients at risk for adverse outcomes is an important goal in preventative medicine and care delivery. However, these associations are often based on complex statistics, and the degree to which a neighborhood or a patient is disadvantaged can be difficult to quantify and may not be readily available in clinical practice.

The Area Deprivation Index (ADI) is a publicly available metric that quantifies neighborhood disadvantage. The ADI is a composite measure that assigns a percentile capturing education, income, employment, and housing characteristics. ¹⁰ A higher ADI indicates a more disadvantaged neighborhood and accounts for economic disparities even within the same zip code. ADI has been extensively studied and is a validated score demonstrated to be an independent predictor of worse health outcomes across multiple medical and surgical specialties. ^{11 –14} However, ADI has not yet been evaluated in relation to urolithiasis or factors affecting urolithiasis.

This study aimed to characterize relationships between ADI and urolithiasis risk factors. We hypothesize that neighborhood socioeconomic disadvantage assessed via ADI is associated with risk factors for recurrent stone disease.

Materials and Methods

After obtaining Institutional Review Board approval, we completed a retrospective review of patients with urolithiasis who required surgical management with percutaneous nephrolithotomy (PCNL) from January 1, 2017, to December 31, 2022, at our large multicenter academic tertiary care center. Adult patients were included if they had a diagnosis of urolithiasis and underwent at least one PCNL during the study period. Patients were excluded from analysis if they were under 18 years of age at the time of the procedure or had a diagnosis of upper tract urothelial carcinoma. All patients were offered the option to complete 24hU testing, as per guidelines and routine protocol at our institution. The final decision to complete 24hU testing varied and was based in part on shared decision-making with the patient. Individual medical records were abstracted for demographic information; past medical history; medications including thiazide diuretics, alkalinization agents, allopurinol, and topiramate; and laboratory results including 24hU parameters and stone composition. In the setting of multiple 24hU studies, the value furthest outside the normal reference range was used for analysis.

ADI is derived from 17 U.S. census variables (Supplementary Table S1) integrated to generate a raw ADI score at the census block group level (600–3000 people). ADI scores for each block are ranked against all block groups nationally to provide a percentile rank from 1 to 100 to serve as a validated metric for neighborhood disadvantage. Complete details of the ADI construction can be found at the Neighborhood Atlas® website (https://www.neighborhoodatlas.medicine.wisc.edu/) by the University of Wisconsin, School of Medicine and Public Health. ⁹ ADI was obtained from our institution’s previously constructed Residency Location History Database, in which patient residential addresses were geocoded to nine-digit zip codes and linked to corresponding national ADI percentile rankings from 1 to 100, then reported in the electronic medical record (EMR). ADI ranks were stratified with the lowest quartile (scores 1–25) representing the least, and the top quartile (76–100) the most disadvantaged.

Results are reported as means and standard deviations, whereas counts and percentages were used to describe categorical variables. Groups defined by ADI quartiles (Q1–Q4) were compared using one-way analysis of variance with Bonferroni post hoc adjustment, Pearson’s chi-square test, and Fisher’s exact tests. Univariable logistic regression and multivariable linear regression were performed to compare 24hU parameters and stone analysis between ADI quartiles, with and without adjustment for sex, and reported as estimates and odds ratios with a 95% confidence interval (CI). A p-value at or below 0.05 was considered statistically significant. SAS 9.4 software (SAS Institute, Cary, NC) was used to perform the analysis.

Results

A total of 1859 adult patients who underwent PCNL for stone disease during the study period comprised the study cohort, with 154 (8.3%) in ADI Q1, 425 (22.9%) patients in Q2, 658 (35.4%) patients in Q3, and 622 (33.5%) patients in Q4.

The distribution of age, sex, race, and body mass index (BMI) was different across ADI quartiles (Table 1). The mean age of patients was 57.9 ± 14.2 years. Patients in the top, most disadvantaged quartile were significantly younger at the time of PCNL compared with all other quartiles (Q1: 60.3 ± 13.7, Q2: 59.2 ± 13.6, Q3: 58.6 ± 13.9, Q4: 55.9 ± 14.7 years, p < .001). There was a higher percentage of females (Q1: 42.2% vs Q4: 55.3%, p = 0.032) and higher mean BMI (Q1: 29.6 ± 6.9 vs Q4: 32.2 ± 8.8 kg/m², p < 0.001) in the most disadvantaged quartile compared with the least disadvantaged quartile. The top quartile also consisted of significantly more black patients compared with all other quartiles (Q1: 3.9%, Q2: 2.6%, Q3: 2.9%, vs Q4: 16.2%, p < 0.001). Patients in the least disadvantaged quartile were significantly more likely to undergo 24hU testing, with 63.6% completing a collection, compared with only 49.9%, 47.9%, and 44.2% in Q2, Q3, and Q4, respectively (p < 0.001). There was no difference in completed nephrology referrals between quartiles.

Among all patients who underwent PCNL, 1811 patients had a stone analysis reported. Mixed calcium and calcium phosphate stone compositions were both associated with ADI quartiles (Table 2). Calcium phosphate stones were more common in the most disadvantaged quartile (p = 0.008), whereas mixed calcium stones were more prevalent in the least disadvantaged quartile (p = 0.027). The odds of having a struvite stone in Q3 and Q2 compared with Q1 was 3.3 (95% CI 1.2–9.4, p = 0.022) and 3.1 (95% CI 1.09–9.0, p = 0.034), respectively (Supplementary Table S2A). The likelihood of having struvite composition was not associated with gender or race.

After excluding 54 patients with a 24hU volume <500 mL, under the assumption of an incomplete collection, there were 846 patients, with 95 (11.2%) patients in Q1, 205 (24.2%) in Q2, 291 (34.4%) patients in Q3, and 255 (30.1%) patients in Q4 (Table 3).

In this subgroup analysis, there was no difference in the distribution of age, sex, or insurance status across quartiles. However, in the top vs bottom quartile, mean BMI was higher (Q4: 32.6 ± 9.0 vs Q1: 29.7 ± 6.0, p = 0.002), and type 2 diabetes mellitus was more prevalent (p = .043). There were otherwise no differences in comorbidities across quartiles. Patients with higher ADI were less likely to live within a 12.5-mile radius of the treating hospital (Q4: 34.6% vs Q1: 51.6%, p = 0.01). The mean 24hU volume was associated with ADI, with patients in the most disadvantaged quartile having significantly lower mean 24hU volume than those in the least disadvantaged quartile (Q4: 1788.8 ± 782.3 vs Q1: 2086.1 ± 788.1 mL, p = 0.009), and 65.9% of patients in Q4 vs 46.3% of patients in Q1 meeting a cutoff value of <2000 mL for low 24hU volume. ADI was also associated with mean 24hU citrate values (p = 0.023). In the top quartile, 63.4% of patients met criteria for hypocitraturia, vs only 57.2%, 55.5%, and 41.6% of patients in Q3, Q2, and Q1, respectively (p = 0.006). The odds of low urine volume and hypocitraturia in patients in Q4 was 2.2 (95% CI 1.4–3.6, p < 0.001) and 2.4 (95% CI 1.5–4.0, p < 0.001) times that of patients in Q1, respectively (Supplementary Table S2B). There was no significant difference in mean 24hU calcium, oxalate, uric acid, sodium, supersaturations, or any other parameter between quartiles. There was a trend toward decreasing 24hU potassium excretion with increasing ADI quartiles, but this did not reach statistical significance (p = 0.058). However, 24hU potassium was significantly lower for patients with hypocitraturia compared with those with normal citrate excretion (63.4 vs 48.9 mg; p < 0.001; Supplementary Table S6).

On multivariable linear regression adjusting for sex and race, these findings remained consistent, except for 24hU potassium, which was significantly lower in Q4 vs Q1 (p = 0.013) in the race adjusted modeling only (Supplementary Table S3).

No difference was found in the prescription of stone prevention medications among ADI quartiles although the indication for a medication was not always clearly described.

There were no differences in 24hU values according to race; however, white patients were more likely to meet the criteria for hyperuricosuria (12.2% vs 5.8%; Supplementary Table S4). Patients with private insurance were significantly more likely to have hyperoxaluria compared with those with Medicaid (41.6% vs 26%; Supplementary Table S8). Otherwise, insurance status was not associated with other 24hU abnormalities. Hyperuricosuria was more prevalent among patients residing >50 miles from the treating hospital compared with those <12.5 miles. However, patients with shorter distance to care were more likely to have high urine sodium (Supplementary Table S7).

Discussion

In this study, we identified differences in patient demographics and the presence of risk factors for urolithiasis based on ADI. Stone formers from disadvantaged neighborhoods required PCNL at a younger age and were also more likely to be female. This observation has also been demonstrated by other groups. For example, Herrick et al. showed that stone formers with state-assisted insurance were more likely to be younger females. ¹⁵ Similarly, Stern et al. and Song et al. reported that younger age and female sex were independent predictors of lower Wisconsin Stone Quality of Life Questionnaire (WISQOL) scores, which may be indicative of underlying factors contributing to the disparate disease presentation. ^16,17 The differences in risk profile observed between ADI quartiles in our study remained consistent even after adjusting for sex. Furthermore, ADI predicted risk factors not identified by insurance status or distance to care, suggesting a strong role of the composite captured by the ADI.

In our cohort, patients residing in disadvantaged neighborhoods had lower mean 24hU volumes and a higher prevalence of hypocitraturia compared with patients from nondisadvantaged neighborhoods. This is critical as fluid intake is widely recognized as a key modifiable risk factor for kidney stone prevention. In a study involving an underserved patient population, 63.9% of patients had a low 24hU volume (<2000 mL), with female gender and insurance status identified as independent predictors of low urine volume on multivariable analysis. ¹⁸ Notably, almost 40% of patients with a low volume on initial 24hU collection were able to improve their urine volume on a subsequent collection, suggesting that this is a correctable risk factor in this population.

The higher prevalence of hypocitraturia in the more disadvantaged quartiles may in part, be attributed to the higher average BMI and higher prevalence of type 2 diabetes, which are both factors with strongly established links to low urinary citrate excretion. ¹⁹ On the contrary, poor SES has also been consistently associated with lower urinary citrate levels. In a single-center study, a higher distressed community index (DCI) was associated with a lower 24hU citrate and potassium, which persisted after multivariate adjustment for age, sex, race, and comorbidities. ²⁰ The authors hypothesized that the observed difference may potentially arise from reduced access to alkali-rich foods, possibly because of restricted food availability as seen in food deserts. In our data, the association between ADI and 24hU potassium excretion did not reach statistical significance; however, hypocitraturia was correlated with lower potassium levels. The latter finding may support decreased fruit/vegetable intake as an explanation for hypocitraturia in our population since 24hU potassium excretion is used as an index for fruit/vegetable consumption. ²¹

ADI did not influence 24hU sodium in our study. However, SES has been linked to significantly higher urinary sodium levels in other studies, for example, by insurance status as mentioned above. ¹⁵ Crivelli et al. used Canadian tax filer data to develop an SES index and also reported a significant inverse correlation between SES and 24hU sodium (p = 0.0002). ²² It is unclear why this association was not observed in our cohort, but urinary sodium excretion may be related to numerous dietary, lifestyle, and environmental factors, as well as chronic diseases and related medication use not captured by our dataset. ²³

We also found that stone composition varied across ADI quartiles. Other groups have observed that patients at more SES disadvantage were more likely to form calcium phosphate stones and had a significantly higher 24hU pH. ¹⁵ However, in our cohort, the higher rate of calcium phosphate stone composition in patients from more disadvantaged neighborhoods was not attributable to a higher 24hU pH, as there was no difference between quartiles. The higher rate of calcium phosphate stones also did not appear to be mediated by 24hU magnesium, which was similar across quartiles. Further studies are needed to better lineate the factors driving the higher prevalence of calcium phosphate stones observed.

Stone disease imposes a significant burden on patients’ health-related quality of life (HRQOL) because of a variety of factors. ²⁴ Several studies have suggested that socioeconomically disadvantaged patients are disproportionately affected by this. For example, one study that found patients from communities in the top social vulnerability index (SVI) quartile had significantly lower WISQOL scores across all domains (69.1 vs 77.2; p = 0.001). ¹⁷ Another study similarly linked stone-specific financial challenges to poorer WISQOL scores (p < 0.001). ²⁵ Possible etiologies proposed for this include travel burden, concerns over taking time off and loss of income, increased disease burden, decreased healthcare use, and delays in care. ²⁶ The authors specifically found that patients with active stone symptoms report worse HRQOL with increased distance to their treatment site. Future work is warranted to address how treatment-related costs may further perpetuate SES disparities in the management of stone disease.

With respect to differences between racial groups, black vs white stone formers have also been found to have a lower 24hU volume, as well as lower citrate, potassium, and magnesium excretion, potentially reflecting differences in dietary intake in addition to genetic factors. ²⁷ Although our study population did have a significantly different racial distribution across quartiles, ADI was associated with risk factors for recurrent stone disease independent of race on multivariate analysis. Univariate analyses in our cohort identified a higher prevalence of hyperuricosuria in whites, but no other differences in 24hU parameters. Several other studies have similarly reported that socioeconomic disadvantage as measured by ADI may mediate perceived race-driven disparities in health outcomes, including longer length of stay after primary total joint arthroplasty, and worse long-term death after orthopedic trauma. ^28,29 Interventions targeting health disparities based on race and ethnicity may not address disparities because of SES; however, focusing on neighborhood disadvantage could help patients facing socioeconomic challenges while also addressing racial or ethnic inequities.

This study is limited by its retrospective nature. We may not have detected differences in some variables, such as 24hU pH, because only the value furthest outside the normal reference range was used for analysis. Supersaturation was also not included for analysis as data were not available for all patients. There may also be some bias introduced as there was a lower rate of 24hU completion among patients with lower SES.

ADI may offer greater clinical utility compared with other socioeconomic or health inequity metrics because of its easy availability and accessibility in the EMR, inclusivity across all U.S. neighborhoods, regular updates, and rigorous validity testing. ⁹ Although the U.S. ADI has limited applicability for patients with a primary address outside of the United States, it may serve as a framework for developing a similar index in other countries. Future work may be aimed at comparing the ADI to other area-level SES indices, including the DCI and SVI.

Conclusions

Patients from socioeconomically disadvantaged neighborhoods undergoing PCNL for urolithiasis were more likely to be female and black and had a lower 24hU test completion rate. Higher ADI was associated with risk factors for recurrent stone disease, including low urine volume and hypocitraturia. Stone types amenable to medical management, including calcium phosphate, were also more prevalent among higher ADI patients. ADI may serve as a simple point-of-care clinical tool to identify patients at risk for neighborhood SES limiting comprehensive endourologic care. Future work is needed to better understand how neighborhood disadvantage affects individual risk and how this information can be used to tailor clinical practice to meet the needs of this patient population.