The economic burden of uncontrolled gout: how controlling gout reduces cost

Abstract

Aim: To evaluate the burden of uncontrolled gout by examining estimated costs and cost drivers.

Materials and methods: Data from the 2012 and 2013 US National Health and Wellness Survey (NHWS; 2012 NHWS, n = 71,157 and 2013 NHWS, n = 75,000) were utilized in this study. Based on self-reported gout diagnosis and gout symptoms, respondents were categorized into three groups: controlled gout (n = 344), uncontrolled gout (n = 2,215), and non-gout controls (n = 126,360). Chi-square tests and one-way analysis of variance (ANOVAs) were used to assess group differences on work productivity loss, healthcare resource utilization, and costs. Zero-inflated negative binomial regressions were used to assess the burden of uncontrolled gout on total costs after controlling for covariates.

Results: Patients with uncontrolled gout had higher presenteeism, overall work impairment, activity impairment, and number of emergency department visits than those with controlled gout or controls. Overall, uncontrolled gout patients had both higher indirect and total costs compared to patients with controlled gout. After controlling for confounders, those with uncontrolled gout had higher total costs than controlled gout respondents and non-gout controls; there was no significant difference in total costs between patients with controlled gout and non-gout controls.

Limitations: Results were based on cross-sectional, self-reported data, making causal inferences more uncertain. Additionally, sample size was small for controlled-gout respondents. Lastly, sampling weights were not used, thus potentially limiting generalizability.

Conclusion: Gout can be an expensive condition, particularly if it is not properly controlled. This study provides support that controlling symptoms (e.g. flares) can reduce the economic and societal burden of gout. Therefore, more attention needs to be paid to effective management of gout symptoms.

Keywords:

• Burden
• cross-sectional
• direct cost
• gout
• healthcare resource use
• indirect cost
• uncontrolled gout
• work productivity

Introduction

Gout, a chronic disease, is the most common cause of inflammatory arthritis in adults. It exerts a significant degree of human and economic burden globally. This is particularly true in the case of uncontrolled gout, which is associated with pain, impairments in quality-of-life, and substantial economic costs^1,2.

Gout is a urate crystal deposition disease caused by chronic high serum uric acid (sUA) levels (i.e. hyperuricemia), which leads to the deposition of monosodium urate crystals in musculoskeletal structures (e.g. joints), kidneys, and throughout the body, resulting in acute gout flares, tophi, and chronic inflammation^3,4. Gout is estimated to affect 3.9% of US adults, and prevalence is increasing worldwide^5,6.

In a recent comprehensive systematic review examining the economic burden of gout, Rai et al.¹ found that gout patients incur both direct and indirect costs that are substantially greater than those without gout. Notably, the annual costs associated with treatment-refractory gout were particularly high compared with patients with controlled disease and non-gout patients. Previous research has demonstrated higher work productivity loss and activity impairment, and higher medical resource use in those suffering from gout⁷.

Further, Wertheimer et al.⁸ utilized US Census and Bureau of Statistics data to further explore the costs associated with gout. The authors estimated that the incremental cost of care of individuals diagnosed with gout was greater than $3,000 annually, with total annual costs in the US in the billions of dollars.

The American College of Rheumatology guidelines for gout management include a target sUA level of <6 mg/dL in all gout case scenarios, and <5 mg/dL to improve more severe signs and symptoms such as palpable and visible tophi. Additionally, sUA levels should be monitored every 6 months once sUA target level has been achieved and more frequently during titration of urate lowering therapy for optimal management⁹. Despite these recommendations, patients’ sUA levels are not consistently monitored, leading to ongoing elevated sUA levels and gout flares¹⁰.

In a recent analysis of data from the National Health and Nutrition Examination Surveys (NHANES) from 2007–2010, Juraschek et al.¹¹ found that over two-thirds of gout patients receiving urate-lowering therapy (ULT) had sUA levels above target. Further, gout is often accompanied by multiple comorbidities that both complicate management and control, and also incur a further economic burden¹². For example, uncontrolled gout with hyperuricemia is associated with an increased risk of hypertension, independent of known traditional hypertension risk factors¹³.

A US-based study conducted by Edwards et al.¹⁴ examined the incremental burden of gout flares among those with refractory chronic gout disease. Among the 81 refractory gout patients surveyed, the average number of gout flares per year was 8.8, with the majority reporting missing workdays due to their condition and an average loss of 25 days annually. Symptomatic gout can, therefore, be associated with a high degree of work-related impairment. In a further US study of 679 individuals with treatment-refractory gout, Wu et al.¹⁵ reported that those diagnosed with gout reported higher rates of healthcare utilization (emergency room visits, hospitalizations, outpatient visits) than matched non-gout patients. The authors noted that these services resulted in incremental healthcare costs of over $10,000 annually.

Many individuals’ gout and urate levels are poorly managed in a clinical setting, which can result in increased physical (e.g. tophi, gout flares) and economic burden. However, less attention has been paid to the differences in health outcomes between those with controlled gout (as measured by achieving target sUA levels) and those with uncontrolled gout. Despite the need for further insight, there exists a distinct paucity of research examining the health and economic burden in those with controlled vs uncontrolled gout.

The importance of patient follow-up in line with published treatment guidelines is widely acknowledged, however, evidence suggests that many patients have uncontrolled disease. This has both humanistic and economic ramifications, and further research to characterize the economic burden of uncontrolled gout can help guide the development and implementation of enhanced care practices. The current study will provide a unique and important contribution to our understanding of the economic ramifications of a highly prevalent and debilitating illness.

Patients and methods

Sample

This project includes data from the 2012 and 2013 US NHWS (2012 NHWS: n = 71,157; 2013 NHWS: n = 75,000). Respondents who took the survey more than once during the 2-year period were only included once and their most recent responses were used, which resulted in n = 130,089 unique respondents. The NHWS is a self-administered, Internet-based questionnaire from a sample of adults (aged 18 or older). A stratified random sample (with strata by gender, age, and ethnicity) was implemented to ensure that the demographic composition of the sample is aligned to that of the corresponding adult population as measured by the US Census. Several previous publications have favorably compared the NHWS, and some of its prevalence estimates, with other governmental sources^16–18. The NHWS has received approval from Essex Institutional Review Board (IRB) (Lebanon, NJ).

Measures

Gout control status

Respondents were classified into three groups based on self-reported gout diagnosis and gout symptoms: controlled gout (n = 344), uncontrolled gout (n = 2,215), and non-gout controls (n = 126,360). Controlled gout was defined as self-reported sUA ≤6 mg/dL and no flares in the past year. Uncontrolled gout was defined as sUA >6 mg/dL or at least one flare in the past year. Gout patients who did not know their sUA levels, or who did not know if they had experienced flares (n = 1,170), were not included in analyses, as the primary objective was to examine control of gout symptoms.

Demographics and health characteristics

Demographic and health characteristic variables were examined for differences between groups: age, sex, race/ethnicity, household income, body mass index (BMI) category, and Charlson Comorbidity Index (CCI)¹⁹.

Cost drivers

Labor force participation and work productivity loss

Labor force participation was examined via one item by asking the respondent’s employment status (employed full-time, employed part-time, or self-employed). Work productivity loss and activity impairment were measured via the Work Productivity and Activity Impairment-General Health scale (WPAI-GH)²⁰. WPAI-GH measured the following based on the past 7 days: Absenteeism (the percentage of work time missed because of health problems), Presenteeism (the percentage of productivity impairment experienced while at work because of health problems), Overall Work Impairment (an overall impairment estimate that is a combination of absenteeism and presenteeism), and Activity Impairment (the percentage of impairment in daily activities because of health problems). Only respondents who reported full-time or part-time employment provided data for Absenteeism, Presenteeism, and Overall Work Impairment; all respondents provided data for Activity Impairment.

Healthcare resource use

Respondents were asked how frequently they had visited various forms of healthcare institutions in the previous 6 months including: healthcare provider visits (HCP), emergency department (ED) visits, and hospitalizations. Healthcare resource use items were then doubled to create annual estimates for number of visits for HCP, ED, and hospitalizations.

Estimated direct and indirect healthcare costs

Estimated annual healthcare resource use costs were calculated by extrapolating data from the 2012 Medical Expenditure Panel Survey (MEPS) to apply as unit costs to Healthcare Resource Use variables from the NHWS (Supplementary Table 1). Similar extrapolation was conducted for indirect costs where 2012 wages from the Bureau of Labor Statistics (BLS) were applied to work hours lost for absenteeism and presenteeism from the WPAI-GH (Supplementary Table 2). Direct costs (ED visits, hospitalizations, HCP visits for all respondents) and indirect costs (absenteeism and presenteeism for employed respondents only) were calculated separately and totaled for the main outcome of total costs. All costs are reported in 2012 US dollars.

Analyses

All analyses were carried out using IBM SPSS version 20 (IBM Corp, Armonk, NY). Demographic and patient characteristic differences were examined for the three comparison groups described above. Means and standard deviations (SD) are reported for continuous variables; frequencies and percentages are reported for categorical variables. For categorical variables, Chi-square tests were used to determine significant differences, while analysis of variance was used for continuous variables when comparing groups.

Following initial comparisons, zero-inflated negative binomial models were used to assess the relationship between gout control and total costs. Zero-inflated negative binomial models test the difference in adjusted relative counts (accounting for covariates) of the costs between gout control groups accounting for the likelihood of incurring any costs. Zero-inflated negative binomial models calculate estimates in two stages: (1) it calculates the likelihood that someone will incur any costs, and (2) examines the actual costs, producing estimated costs weighted by the likelihood of incurring a cost. Covariates included were: age, sex, ethnicity, income, BMI category, and CCI.

Results

Demographics and health characteristics

On average, patients with controlled gout were aged 62.6 years (SD = 12.0), the majority were male (87.2%) and white (82.3%), 41.6% had an annual income of $75,000 or more, 48.0% were characterized as obese, and they had a mean CCI score of 0.95 (SD = 1.3). Patients with uncontrolled gout were aged 59.8 years on average (SD = 12.8), were predominantly male (77.7%) and white (77.9%), 27.1% had an annual income of $75,000 or more, 57.3% were obese, and they had a CCI score of 1.20 on average (SD = 1.8). Lastly, non-gout controls were aged 47.6 years on average (SD = 16.7), the majority were female (53.3%) and white (71.3%), 26.2% earned $75,000 or more annually, 31.0% were obese, and they had a mean CCI score of 0.42 (SD = 1.0). When comparing across all three groups, there were statistically significant group differences for all demographic and health characteristic variables (all p < .001; Table 1).

Table 1. Demographics and health characteristics by gout/non-gout and gout control status.

	Gout control						Omnibus Chi-square/one-way ANOVA p-value
	Controlled gout (n = 344)		Uncontrolled gout (n = 2,215)		Non-gout controls (n = 126,360)
	Mean/Column %	SD/Count	Mean/Column %	SD/Count	Mean/Column %	SD/Count
Age of respondent	62.58	12.02	59.79	12.75	47.62	16.66	<.001
Sex
Female	12.8%	44	22.3%	494	53.3%	67,289	<.001
Male	87.2%	300	77.7%	1,721	46.7%	59,071
Race/ethnicity
Non-Hispanic white	82.3%	283	77.9%	1,726	71.3%	90,075	<.001
Non-Hispanic black	3.8%	13	11.2%	248	11.6%	14,687
Hispanic	5.2%	18	4.8%	106	8.8%	11,062
Other ethnicity	8.7%	30	6.1%	135	8.3%	10,536
Annual household income
< $25k	10.5%	36	20.8%	460	19.5%	24,691	<.001
$25k to $50k	22.1%	76	27.4%	608	26.9%	33,944
$50k to $75k	19.2%	66	20.1%	446	19.7%	24,894
$75k or more	41.6%	143	27.1%	601	26.2%	33,138
Decline to Answer	6.7%	23	4.5%	100	7.7%	9,693
BMI category
Underweight	0.3%	1	0.4%	8	2.0%	2,588	<.001
Normal weight	9.9%	34	10.4%	231	32.6%	41,232
Overweight	39.8%	137	30.2%	670	31.7%	40,010
Obese	48.0%	165	57.3%	1,269	31.0%	39,221
Unknown	2.0%	7	1.7%	37	2.6%	3,309
Charlson comorbidity index scoring	0.95	1.28	1.20	1.80	0.42	1.00	<.001

Abbreviations. ANOVA, analysis of variance; BMI, body mass index; SD, standard deviation.

p-values represent the omnibus analyses assessing differences across all three groups.

Initial analyses of outcomes and cost drivers (WPAI-GH and resource use)

For labor force participation, patients with gout (either controlled or uncontrolled) were less likely to participate in the labor force than non-gout controls (p < .001). Patients with uncontrolled gout had higher presenteeism, overall work impairment, and activity impairment than those with controlled gout (p < .05). The only significant difference between controlled gout and non-gout controls was that patients with controlled gout reported higher activity impairment than non-gout controls (p < .05; Table 2).

Table 2. Cost drivers by gout control status.

	Gout control						Omnibus Chi-square p-value
	Controlled gout (n = 344)		Uncontrolled gout (n = 2,215)		Non-gout controls (n = 126,360)
	Column %	Count	Column %	Count	Column %	Count
Participate in the labor force	35.8%a	123	37.7%a	835	53.9%b	68,117	<.001
	Mean	SD	Mean	SD	Mean	SD	Omnibus one-way ANOVA p-value
Work productivity and activity impairment
Absenteeism* (% work missed)	3.18a,b	12.78	5.29a	16.27	3.59b	13.89	.003
Presenteeism* (% impaired while working)	12.88a	20.88	21.57b	25.97	13.43a	22.41	<.001
Overall work impairment* (% overall work impairment)	15.50a	23.78	24.22b	28.81	15.50a	25.31	<.001
Activity impairment (% activity impairment)	27.38a	29.58	38.23b	31.66	22.04c	28.22	<.001
Resource use in the past 12 months
Number of visits to any HCP	11.78a	11.53	11.43a	12.60	7.13b	11.46	<.001
Number of visits to ED	0.31a	0.90	0.80b	2.95	0.42a	2.22	<.001
Number of hospitalizations	0.29a,b	1.14	0.51a	2.51	0.24b	2.01	<.001

*Asked only of those participating in the labor force.

Abbreviations. ANOVA, analysis of variance; ED, emergency department; HCP, healthcare provider; SD, standard deviation.

p-values represent the omnibus analyses assessing differences across all three groups. For pairwise comparisons, refer to subscripts (e.g. a, b, c) attached to means or percentages. Mean values or percentages in the same row that do not share the same subscript (e.g. XXq vs XXb,) are significantly different at p < .05. If they share the same subscript (e.g. XXa vs XXa or XXa,b vs XXa) then they are not significantly different at p < .05.

Regarding healthcare resource use, those with uncontrolled gout had a higher number of ED visits than those with controlled gout or non-gout controls (p < .05). There were no differences between uncontrolled gout and controlled gout for the number of doctor visits or number of hospitalizations. Lastly, there were no differences between controlled gout and controls on number of ED visits or number of hospitalizations; however, controls had fewer doctor visits than those with controlled gout (p < .05; Table 2).

Overall, the burden of uncontrolled gout was evidenced by higher total costs than controlled gout and non-gout controls (p < .05). Further breaking down cost components, those with uncontrolled gout had higher costs associated with ED visits, hospitalizations, presenteeism, and absenteeism than those with controlled gout (p < .05). Additionally, those with uncontrolled gout had higher costs associated with all cost variables compared with non-gout controls (p < .05). Lastly, those with controlled gout reported higher doctor visit costs, direct costs, and total costs than non-gout controls (p < .05; Table 3).

Table 3. Annual direct and indirect costs by gout control status.

	Gout control						One-way ANOVA p-value
	Controlled gout (n = 344)		Uncontrolled gout (n = 2,215)		Non-gout controls (n = 126,360)
	Mean	SD	Mean	SD	Mean	SD
Total direct costs	$23,574a	$27,655	$27,074a	$48,727	$13,412b	$37,682	<.001
Cost of HCP visits	$18,081a	$17,719	$16,447a	$18,057	$8,809b	$14,695	<.001
Cost of ED visits	$459a	$1,315	$1,227b	$4,565	$636a	$3,371	<.001
Cost of hospitalizations	$5,033a	$17,529	$9,400b	$39,347	$3,967a	$30,335	<.001
Total indirect costs	$5,622a	$9,913	$9,557b	$13,680	$5,174a	$10,049	<.001
Cost of absenteeism	$1,024a	$5,627	$2,120b	$7,843	$1,159a	$5,428	<.001
Cost of presenteeism	$4,606a	$8,259	$7,436b	$10,504	$4,016a	$7,805	<.001
Total costs	$25,584a	$28,394	$30,672b	$49,670	$16,197c	$38,980	<.001

ANOVA, analysis of variance; ED, emergency department; HCP, healthcare provider; SD, standard deviation.

p-values represent the omnibus analyses assessing differences across all three groups. For pairwise comparisons, refer to subscripts (e.g. a, b, c) attached to means or percentages. Mean values or percentages in the same row that do not share the same subscript (e.g. XXq vs XXb,) are significantly different at p < .05. If they share the same subscript (e.g. XXa vs XXa or XXa,b vs XXa) then they are not significantly different at p < .05.

Multivariable analyses

To examine in more detail the unique burden of uncontrolled gout on total costs, a zero-inflated negative binomial generalized linear model was conducted to see how total costs differ by gout status after controlling for demographic and health characteristics.

After controlling for relevant confounders (age, sex, ethnicity, income, BMI category, and CCI), results showed that those with uncontrolled gout had significantly higher total costs than those with controlled gout and non-gout controls. The difference between non-gout controls and controlled gout was not statistically significant (Table 4).

Table 4. Total costs by gout control status.

	Gout control
	Controlled gout (n = 344)					Uncontrolled gout (n = 2,215)			Non-gout controls (n = 126,360) Estimated mean (SE)
	Estimated mean (SE)	Coefficient vs non-gout controls	Coefficient p-value vs non-gout controls	Coefficient vs uncontrolled gout	Coefficient p-value vs uncontrolled gout	Estimated mean (SE)	Coefficient vs non-gout controls	Coefficient p-value vs non-gout controls
Total cost	$19,166 ($1,121)	0.069	.226	–0.124	.042	$21,191 ($503)	0.193	<.001	$15,819 ($57)

Abbreviations. SE, standard error.

Analysis controlled for confounders from a multivariable regression analysis using a zero-inflated negative binomial model. Covariates included: age, sex, ethnicity, income, body mass index category, and Charlson Comorbidity Index.

Discussion

Gout is a chronic disease that exerts a significant degree of human and economic burden globally. This is particularly true for uncontrolled gout, which is associated with pain and impairments in quality-of-life. Further, uncontrolled symptoms can lead to substantial incremental economic costs, including ED visits, hospitalizations, and work productivity losses, many of which could be reduced through effective symptom control among this population of sufferers.

A number of notable differences emerged in the demographic characteristics of gout patients with uncontrolled disease compared to those with controlled disease, as well as those without gout (Table 1). This included significant differences based on gender, age, race/ethnicity, BMI, annual income, and comorbid disease.

The current study established that individuals with uncontrolled gout generally reported significantly greater healthcare burden and incurred higher economic costs when compared to those with controlled gout. This is in line with previous studies that have established the broader economic consequences of gout¹, as well as the incremental complexity and burden associated with uncontrolled and treatment refractory gout^13,14,16. This included greater work-related impairment (e.g. presenteeism), as well greater health resource use (e.g. ED visits, hospitalizations). Importantly, few significant or substantial differences emerged between patients reporting controlled gout symptomatology and the control group of those without gout.

In the multivariable analyses, results showed that those with uncontrolled gout had significantly higher total costs than those with controlled gout and non-gout controls, after controlling for potential confounding factors. However, the difference was smaller and not significant between non-gout controls and patients with controlled gout.

Critically, the results of this study suggest that the effective control of gout symptoms can serve to reduce the cost burden associated with gout. In other words, individuals with controlled gout reported similar outcomes in many cases as non-gout controls, and, importantly, these two groups remained distinct from those with uncontrolled gout, accounting for confounding factors. This emphasizes the critical importance of promoting effective management of gout in the clinical setting, which could promote not only significant cost-savings, but also enhanced health and quality-of-life outcomes among sufferers.

Limitations and future directions

The novel findings established in the current study must be considered in light of its limitations. The study utilized a large-scale survey approach, and thus relied upon self-reported symptomatology and diagnoses that were not possible to verify clinically. Whereas this is important to note, past survey research has established good comparability between self-reported prevalence and clinical data^16–18. Further, the cross-sectional nature of the study makes causal inference more difficult.

Whereas significant efforts were made to recruit a large and representative sample of subjects, the sample size of some study groups was limited. This was largely due to many gout patients being unable to report their current sUA levels or recent flares, thereby excluding themselves from the study. This may be reflective of patients’ lack of awareness of their symptoms and symptom management, and thus further reinforces the importance of raising awareness of this clinical care issue. Finally, sampling weights were not used in analyses, and thus results may not be possible to directly extrapolate to the adult US population; however, overall the NHWS does use stratified sampling to ensure demographic composition is similar to that of the US population.

The results of the current study provide a preliminary but important foundation for future research exploring the health and economic burden of uncontrolled gout. Future research should seek to build on these findings and examine effective ways by which to develop and implement clinical care practices that will result in higher rates of controlled disease.

Conclusions

Gout is a burdensome but clinically manageable disease condition. The current study suggests that the effective management of sUA levels and flares can reduce the cost burden associated with gout, and emphasizes the critical importance of promoting effective management of gout in the clinical setting. Increased attention to this facet of care could be associated with not only significant cost-savings, but also enhanced health and quality-of-life outcomes among sufferers.

Transparency

Declaration of funding

This study was funded by AstraZeneca, but the study was conducted at Kantar Health. Sponsors of the study assisted with manuscript development and are co-authors.

Declaration of financial/other relationships

AstraZeneca paid Kantar Health for access to NHWS data, analyses, and manuscript preparation. NF was a full-time employee at Kantar Health at the time of the study and manuscript development. SB and RM were employees of Ardea Bioscience, Inc., a member of the AstraZeneca Group, at the time of the study and manuscript development. AK and JN are full-time employees of AstraZeneca. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

Errol J. Philip contributed to manuscript preparation and provided editorial support and is a paid consultant of Kantar Health. Editorial support was also provided by Albert Balkiewicz, MSc, of PAREXEL, which was funded by AstraZeneca.

Additional information

Notes on contributors

Natalia M. Flores

All authors fully contributed to the design and conduct of the analysis, review, access, and interpretation of data, and critical drafting, review, and approval for submission of the manuscript.

Javier Nuevo

All authors fully contributed to the design and conduct of the analysis, review, access, and interpretation of data, and critical drafting, review, and approval for submission of the manuscript.

Alyssa B. Klein

All authors fully contributed to the design and conduct of the analysis, review, access, and interpretation of data, and critical drafting, review, and approval for submission of the manuscript.

Scott Baumgartner

All authors fully contributed to the design and conduct of the analysis, review, access, and interpretation of data, and critical drafting, review, and approval for submission of the manuscript.

Robert Morlock

All authors fully contributed to the design and conduct of the analysis, review, access, and interpretation of data, and critical drafting, review, and approval for submission of the manuscript.