- Email: [email protected]
Clinical Characteristics, Health Care Utilization and Costs Among Men with Primary or Secondary Hypogonadism in a US Commercially Insured Population
ENDOCRINE
Clinical Characteristics, Health Care Utilization and Costs Among Men with Primary or Secondary Hypogonadism in a US Commercially Insured Population Michael Grabner, PhD,1 Amit Bodhani, MS,2 Nikhil Khandelwal, PhD,2 Swetha Palli, MS,3 Nicole Bonine, PhD,4 and Mohit Khera, MD5
ABSTRACT
Introduction: Hypogonadism is broadly associated with increases in chronic comorbid conditions and health care costs. Little is known about the specific impact of primary and secondary hypogonadism on health care costs. Aim: To characterize the health care cost and utilization burden of primary and secondary hypogonadism in a population of US men with commercial insurance. Methods: Newly diagnosed patients with International Classification of Diseases, Ninth Revision, Clinical Modification codes associated with specific medical conditions known to have a high prevalence of testosterone deficiency (ie, relating to primary or secondary hypogonadism) or who had fills for testosterone replacement therapy from January 1, 2007 through April 30, 2013 were identified in administrative claims data from the HealthCore Integrated Research Database. A cohort of patients without hypogonadism was matched on demographics and comorbidities. The matched hypogonadism and non-hypogonadism cohorts (n ¼ 5,777 in each cohort) were compared during a 12-month follow-up period. Main Outcome Measures: Direct health care expenditures and utilization were assessed for all causes and for hypogonadism-related claims. Costs included out-of-pocket patient expenditures and those paid by the insurer. Results: Hypogonadism and matched non-hypogonadism cohorts were similar in demographics (mean age ¼ 50 years) and diagnosed comorbid conditions in the 12 months preceding the index date. In the year after the index date, mean allcause expenditures for patients with hypogonadism increased by 62% (from $5,425 to $8,813) compared with 25% for the matched controls (from $4,786 to $5,992; P < .01 for follow-up difference between groups). Approximately 16% of total mean costs ($1,377), primarily outpatient and pharmacy costs, were identifiable as related to hypogonadism. Conclusion: These data from a population of US men with commercial insurance coverage showed a greater resource use burden for patients with primary and secondary hypogonadism compared with similar patients without hypogonadism. Additional management might be required to address unmet need and decrease the cost burden for patients with hypogonadism. J Sex Med 2017;14:88e97. Copyright 2016, The Authors. Published by Elsevier Inc. on behalf of the International Society for Sexual Medicine. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Key Words: Hypogonadism; Testosterone Deficiency; Retrospective Study; Health Care Utilization; Costs and Cost Analysis
INTRODUCTION Male hypogonadism has been defined as men with persistent testosterone deficiency and, according to Endocrine Received August 24, 2016. Accepted October 29, 2016. 1
HealthCore, Inc, Wilmington, DE, USA;
2
AbbVie Inc, Chicago, IL, USA;
3
CTI Clinical Trial and Consulting Services, Cincinnati, OH, USA;
4 5
Society guidelines, might be associated with symptoms such as decreased libido, sexual and erectile dysfunction, fatigue, lethargy, and mood disturbances.1,2 Decreased lean muscle
Copyright ª 2016, The Authors. Published by Elsevier Inc. on behalf of the International Society for Sexual Medicine. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). http://dx.doi.org/10.1016/j.jsxm.2016.10.012
OptumRx, Irvine, CA, USA;
Baylor College of Medicine Medical Center, Houston, TX, USA
88
J Sex Med 2017;14:88e97
89
Cost Burden of Primary or Secondary Hypogonadism
mass and bone density are common and have been associated with fractures and frailty.2 Testosterone deficiency also has been associated with increased chronic comorbidities and health care costs.3,4 Specific causes of primary hypogonadism include genetic conditions (eg, Klinefelter syndrome), anatomic defects, infection, tumor, injury, iatrogenic causes (surgery or certain medications), and/or alcohol abuse. Secondary hypogonadism (low levels or inappropriately normal levels of follicle-stimulating hormone and luteinizing hormone in the presence of low testosterone) also can be caused by inherited conditions (eg, Kallmann syndrome), pituitary disorders, or hypothalamic lesions or disorders.5 The decrease of testosterone with increasing age is related to testicular and hypothalamic-pituitary function and is referred to as age-related hypogonadism or late-onset hypogonadism.1,6 There is limited published research estimating the specific impact of primary and secondary hypogonadism on health care costs. Understanding the relation of hypogonadism to costs over time requires isolating the impact in the context of other comorbidities. We attempted to advance this understanding by examining the costs of patients with primary and secondary hypogonadism longitudinally in a commercially insured population compared with matched controls.
AIMS The purpose of the present study was to estimate the direct cost burden of primary and secondary hypogonadism to health care payers and patients compared with a matched control group without hypogonadism.
METHODS We conducted a retrospective analysis using a US database of commercial health plan claims. Because the goal of the retrospective study was to include and summarize data for all eligible patients in the database, a predetermined sample size was not calculated. Throughout this study, researchers’ access was limited to data removed of all identifiers to ensure confidentiality. No study patients were contacted and informed consent was not obtained. An institutional review board did not review the study because only data in the format of a limited dataset were accessed and HealthCore maintains data use agreements with the covered entities in compliance with the Health Insurance Portability and Accountability Act of 1996.
Data Source All study data were derived from the HealthCore Integrated Research Database (HIRD), a repository of fully adjudicated medical and pharmacy claims on more than 37 million health plan members from across the United States. The commercial J Sex Med 2017;14:88e97
health plans included in the HIRD encompass health maintenance organizations, point of service, preferred provider organizations, indemnity plans, and Medicare Advantage plans. Dates queried spanned the period from January 1, 2006 through April 30, 2014. To allow for a 12-month pre-index period and a follow-up time of 12 months, the study intake period (containing the index date) was set at January 1, 2007 to April 30, 2013.
Patient Selection International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes and prescriptions for testosterone replacement therapy (TRT) were used to identify patients with hypogonadism. Adult men (18 years old) with commercial insurance coverage (employer-sponsored or Medicare Advantage) were included if they had at least two medical claims for conditions indicating hypogonadism (of any type; Appendix A presents a list of codes), or at least two pharmacy claims for TRT, or one claim for TRT and one claim for hypogonadism. The index date for patients with hypogonadism was defined as the date when patients first met at least one criterion for hypogonadism (from medical or pharmacy claims). For patients without hypogonadism, the index date was defined as the earliest date during a control patient’s enrollment period that created a successful match. The follow-up time was 12 months for all patients. Patients who died or lost coverage during the follow-up period were not included in the analysis because continuous enrollment was required for 12 months before and after the index date. A matched cohort of patients without hypogonadism, otherwise similar in demographic and clinical characteristics, was identified for comparison. A 1:1 matching algorithm was constructed to incorporate demographic characteristics, selected comorbidities, and baseline level of resource use (Appendix B presents a list of all matching criteria). The comorbidities for matching were frequently observed in the hypogonadism population and/or considered important in this setting based on a priori clinical input. Final criteria were selected with the goal of obtaining a close match on key confounding variables while preserving a large portion of the initial hypogonadism cohort. After creation of the matched cohorts, all patients with claims whose ICD-9-CM codes suggested the presence of age-related hypogonadism were excluded before the final analysis.
Analyses Descriptive statistics included means (SDs) for continuous data and absolute and relative frequencies for categorical data. For comparisons between the hypogonadism and nonhypogonadism groups, we used c2 tests for categorical variables, t-tests for normally distributed continuous variables, and Wilcoxon rank-sum tests for non-normally distributed continuous variables. All tests were two-sided using an a value of
90
Grabner et al
Figure 1. Patient attrition.
0.05. No adjustments for multiple comparisons were performed. Because the large samples available in administrative databases can lead to relatively trivial differences in magnitude being characterized as highly significant statistically, standardized mean differences (SMDs) were calculated between the cohorts to assist in interpretation of effect sizes.7 SMDs in absolute value of more than 10 were considered to have potential clinical significance.8 Analyses were performed using Excel 2007 (Microsoft, Redmond, WA, USA) and SAS 9.4 (SAS Institute, Cary, NC, USA).
MAIN OUTCOME MEASURES Resource utilization (hospital and emergency department admissions, outpatient visits including office visits and laboratory tests,
and pharmacy fills) and costs were assessed from medical and pharmacy claims during the 12-month baseline period and the 12month follow-up period. Costs were adjusted to 2014 US dollars using the medical care component of the Consumer Price Index.9 All-cause health care costs were reported in aggregate and segmented across places of service and included plan-paid costs and patient out-of-pocket expenditures. Resource utilization and costs were considered related to hypogonadism if the associated claims contained an ICD-9-CM diagnosis code indicative of hypogonadism or a procedure code for a testosterone panel or were pharmacy claims for TRT. All diagnosis fields, at any time during a hospital stay, were used to categorize inpatient hospitalizations. By definition, the non-hypogonadism cohort had no hypogonadism-related costs. Costs and utilization for the 12-month period after hypogonadism diagnosis were the main outcome metrics. J Sex Med 2017;14:88e97
91
Cost Burden of Primary or Secondary Hypogonadism
Table 1. Baseline demographics of study population*
Age (y), mean (SD)† Age (y), % 18e44 45e64 65 Type of health plan, %† HMO PPO Other Region, %† Northeast Midwest South West Index year, %† 2007e2008 2009e2010 2011e2012 2013 (JanuaryeApril) Treating MD specialty, % Primary care Urology Endocrinology Other or missing‡
Hypogonadism cohort before matching (n ¼ 9,463)
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
51 (11.6)
50 (11.1)
49 (11.1)
28.5 63.2 8.3
29.8 64.1 6.1
31.2 62.5 6.3
26.0 66.6 7.5
25.2 68.1 6.8
25.2 68.1 6.8
15.0 34.4 22.9 27.7
14.1 35.2 22.2 28.6
14.1 35.2 22.2 28.6
29.0 31.9 35.3 3.8
29.8 31.7 34.9 3.5
29.8 31.7 34.9 3.5
61.1 9.6 5.7 23.6
61.9 9.5 5.5 23.1
N/A N/A N/A N/A
HMO ¼ health maintenance organization; N/A ¼ not applicable; PPO ¼ preferred provider organization. *Categories might not total to 100% because of rounding. No statistically significant differences were noted between matched groups (P > .05) for demographic variables, and all absolute standardized mean differences were less than 10. No significance testing was conducted between the full hypogonadism cohort before matching and the matched groups. † Indicates a patient characteristic used in matching patients with and without hypogonadism (Appendix A). ‡ Includes other specialties (eg, geriatrics) and non-physicians.
RESULTS A total of 9,463 patients with claims for primary or secondary hypogonadism or TRT and who fulfilled all inclusion and exclusion criteria were included in the analysis (Figure 1). The mean age at index date was 51 years; 61% patients were treated in a primary care setting; and the largest geographic representation was in the Midwest (Table 1). Most patients received coverage under preferred provider organization health plans. After matching the patients with hypogonadism to a control cohort of patients without hypogonadism, 5,777 patients remained in each cohort. Overall, demographic characteristics in the full hypogonadism-TRT cohort were similar to those in the subset of matched patients with hypogonadism. The matched cohort was healthier than the full hypogonadism-TRT group, with fewer comorbid medical conditions (Table 2, Appendix C) and a lower mean Quan-Charlson comorbidity index score. As intended, the matched hypogonadism and nonhypogonadism cohorts were similar in patient characteristics. Hypertension and hyperlipidemia were common (46.3% and J Sex Med 2017;14:88e97
53.6%, respectively), and approximately 18% of patients had diabetes. Differences were seen in several variables not included in the list of matching requirements; for instance, baseline medication use, including mental health medication, was more common in the hypogonadism group than in the matched non-hypogonadism control group. However, in all cases, the absolute SMDs were below the threshold of 10. Meaningful clinical differences between the groups were not apparent at baseline for any of the demographic or clinical variables evaluated (Table 2). Health care utilization for patients with and without hypogonadism, while well matched at baseline, diverged in the year after the index date (Table 3). Patients with hypogonadism had a mean of 5.8 physician office visits and 8.4 visits to other outpatient services in the year after their diagnosis compared with 4.3 and 6.1, respectively, for patients without hypogonadism (P < .0001; SMD ¼ 28.8 and 22.8). Slightly more patients with than without hypogonadism were admitted to the hospital (7.9% vs 6.6%), although the magnitude of the difference was not large enough to be considered clinically relevant (SMD < 10). Patients with hypogonadism had an average of 28.2 pharmacy fills (12% of
92
Grabner et al
Table 2. Baseline comorbidities and medications (within 12 months before index date)
Comorbidities, % Hypertension* Hyperlipidemia* Coronary heart disease* Cerebrovascular disease Myocardial infarction Diabetes (mild to moderate)* Diabetes (with complication) Renal failure* Obesity* Obstructive sleep apnea* Asthma† Chronic pulmonary disease Osteoarthritis Osteoporosis† Depression* Anxiety† Fatigue* HIV-AIDS Malignancy (except skin neoplasms) QCI, mean (SD)* QCI, % 0e1 2 3 Baseline medications Antihypertensives† Antihyperlipidemics† Antidiabetics† Mental health medications†
Hypogonadism cohort before matching (n ¼ 9,463)
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
43.3 47.4 12.4 4.0 1.6 19.7 3.4 12.4 7.2 7.5 12.7 9.6 14.3 3.5 24.8 26.9 30.9 2.0 5.5 0.9 (1.6)
46.3 53.6 7.3 2.0 0.7 16.2 1.6 6.8 3.9 3.9 10.5 6.7 10.7 2.4 19.1 21.3 25.5 0.9 1.6 0.5 (0.9)
46.3 53.6 7.3 1.7 0.8 15.7 1.5 6.8 3.9 3.9 9.0 7.4 10.7 1.2 19.1 19.4 25.5 0.8 1.6 0.5 (0.9)
60.7 20.5 18.7
69.9 21.1 9.0
70.9 20.7 8.4
35.1 37.7 15.9 40.2
33.0 35.9 13.1 33.6
30.1 31.4 11.6 29.0
QCI ¼ Quan-Charlson Comorbidity Index (modified from Charlson Comorbidity Index for use in administrative databases).10 *Indicates a patient characteristic used in matching patients with and without hypogonadism (Appendix A). † P < .05 between matched groups, with absolute standardized mean differences less than 10 in all cases. No significance testing was conducted between the full hypogonadism cohort before matching and the matched groups. See Appendix C for individual P values and standardized mean differences.
which were related to hypogonadism) compared with 17.5 for patients without hypogonadism (P < .0001; SMD ¼ 46.4). In the year after hypogonadism diagnosis, mean all-cause expenditures for patients with hypogonadism increased by 62% (from $5,425 to $8,813) compared with 25% for matched controls (from $4,786 to $5,992); costs in the hypogonadism cohort were significantly higher compared with those in the matched control cohort ($8,813 vs $5,992; P < .001; SMD ¼ 15.0; Figure 2). Outpatient visits and pharmacy costs were the largest contributors to total cost and to cost increases. Approximately 15% of total costs, primarily outpatient and pharmacy costs, were identifiable as related to hypogonadism (mean ¼ $1,377; median ¼ $718 per patient per year; costs were considered related to hypogonadism if the associated claims contained an ICD-9-CM diagnosis code indicative of hypogonadism or a procedure code for a testosterone panel or were pharmacy claims for TRT).
DISCUSSION This retrospective study from a commercial US health plan describes a population of more than 5,000 men with newly diagnosed primary or secondary hypogonadism. We found that patients with primary and secondary hypogonadism had increased health care utilization and cost during the year after their diagnosis compared with matched controls without diagnosed hypogonadism but with a similar baseline comorbidity burden. These results are consistent with those of a previous study using retrospective data, also in the United States, but from a different source.11 Kaltenboeck et al11 examined a prevalent, rather than incident, hypogonadism cohort (including primary, secondary, and age-related etiologies) and used risk adjustment rather than direct matching to estimate control group costs. Annual risk-adjusted all-cause direct costs in that study were J Sex Med 2017;14:88e97
93
Cost Burden of Primary or Secondary Hypogonadism
Table 3. Health care utilization during 12 months after index date
Patients with 1 inpatient hospitalization, n (%)* With 1 hypogonadism-related hospitalization, % Patients with 1 emergency room visit, n (%) With 1 hypogonadism-related visit, % Physician office visits per patient per year, mean (SD)†,‡ Total visits related to hypogonadism, % Other outpatient visits per patient per year, mean (SD)†,‡ Total visits related to hypogonadism, % Pharmacy, number of fills per patient per year, mean (SD)†,‡ Total fills for hypogonadism, %
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
457 8.3 750 1.6 5.8 12.2 10.4 14.6 24.8 12.4
384 N/A 713 N/A 4.3 N/A 6.2 N/A 17.0 N/A
(7.9) (13.0) (5.8) (11.6) (22.7)
(6.6) (12.3) (5.0) (9.9) (20.2)
N/A ¼ not applicable. *Any emergency room visit resulting in hospital admission was counted toward inpatient hospitalizations. † Absolute standardized mean difference greater than 10. ‡ P < .0001 between groups.
$9,291 for hypogonadism (any type) vs $5,248 for nonhypogonadism controls. Of note, because the study by Kaltenboeck et al focused on employed men, the investigators could estimate indirect costs from disability and absenteeism, which added an additional 22% to the risk-adjusted cost difference. Comorbidity profiles for patients with hypogonadism (any type)
Figure 2. All-cause health care costs for 12 months before and after index date (mean; 2014 US dollars). Costs during 1-year follow-up (POST) were significantly higher in the hypogonadism cohort than in the non-hypogonadism cohort ($8,813 vs $5,992; P < .001; standardized mean difference ¼ 15.0). J Sex Med 2017;14:88e97
were comparable to those seen in the present study. A second, more recent, study using administrative claims data came to a similar conclusion regarding cost differences between patients with vs without primary or secondary hypogonadism; in a sample of 7,386 matched patients from January 2006 through June 2014, all-cause total costs ($13,686 vs $7,551) and medical costs ($10,675 vs $5,769) were significantly higher in patients with hypogonadism.12 Other past studies, using different data sources and methods, also have found an increase in utilization and costs for patients with hypogonadism.13,14 Strengths of the present study include the large number of analyzed patients and the creation of a well-defined comparison cohort by matching. Because the intake period began in 2007, all patients began hypogonadism treatment after the 2006 Endocrine Society guidelines became available.15 Tight matching on several patient characteristics was possible given the large initial sample and the detail captured in the administrative claims. The small proportion of patients at least 65 years old in our analysis reflects well on our efforts to focus on a primary and secondary hypogonadism population; nevertheless, external validation of the set of ICD-9-CM codes that was used would be valuable. The matching algorithm was successful in creating cohorts of primary and secondary hypogonadism and non-hypogonadism that were similar at baseline in all demographic and clinical aspects. We incorporated baseline resource utilization and other operational variables (plan type and region) known to affect health care costs in the matching algorithm. The overall pre-match hypogonadism-TRT population was less healthy compared with the more strictly defined matched group, suggesting that resource utilization and costs could be even higher in the overall hypogonadism population compared with the non-hypogonadism population. Limitations of the study are primarily a result of reliance on medical claims data, which lack clinical context and are
94
subject to potential coding, billing, and recording errors and other inconsistencies. The lack of availability of certain clinical data meant we could not replicate current diagnostic criteria for hypogonadism. Specifically, testosterone and gonadotropin levels were not used to determine inclusion in the present analysis because laboratory data were available only for a small subset of patients in this data source. Clinical symptoms reported by patients that did not lead to a billable health care encounter would have been missed. Therefore, the control cohort could include a small percentage of men with low testosterone levels who might not have been diagnosed with hypogonadism. The two cohorts also included patients based on the presence of TRT claims only, without requiring the presence of ICD-9-CM codes for primary or secondary hypogonadism; however, those with ICD-9-CM codes for agerelated hypogonadism were excluded (a small proportion of patients might have had age-related hypogonadism without the presence of a related code). Similarly, important risk factors such as obesity are typically under-reported in claims data. Conversely, the presence of a diagnosis code on a medical claim does not guarantee the presence of a disease, because the diagnosis code might be incorrectly coded or included as a rule-out criterion. We required at least two claims for hypogonadism or TRT on distinct dates to limit this bias. Claims databases also only include patients with specific health care coverage, in this case private commercial health plans. Notably, patients without health insurance or those who are covered under public programs (such as Medicare or Medicaid) are not part of the analyzed population. Therefore, the ability to generalize these results to the entire US population might be limited. Commercially insured individuals are more likely to be currently employed and tend to be healthier than the general population, especially if they continue working. Further research is necessary to determine how primary and secondary hypogonadism contributes to increased health care costs over time. The constellation of symptoms observed in patients with hypogonadism—decreased energy, loss of muscle, depressed mood, fractures, and frailty—can lead to physical inactivity and aggravate the attendant risk of diabetes and cardiovascular disease. This study was limited to 1 year of follow-up after hypogonadism diagnosis, so a full picture of the economic burden of hypogonadism might require longer observation. More research also is needed to investigate the potential benefit of TRT in mitigating the burden of hypogonadism. Correspondence Author: Michael Grabner, PhD, HealthCore, Inc, 123 Justison Street, Suite 200, Wilmington, DE 19801, USA. Tel: 530-400-5978; E-mail: [email protected]
Grabner et al
STATEMENT OF AUTHORSHIP Category 1 (a) Conception and Design Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera (b) Acquisition of Data Michael Grabner; Swetha Palli; Nicole Bonine (c) Analysis and Interpretation of Data Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera Category 2 (a) Drafting the Article Michael Grabner; Swetha Palli; Nicole Bonine (b) Revising It for Intellectual Content Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera Category 3 (a) Final Approval of the Completed Article Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera
REFERENCES 1. Bhasin S, Cunningham GR, Hayes FJ, et al. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2010;95:2536-2559. 2. Petak SM, Nankin HR, Spark RF, et al. American Association of Clinical Endocrinologists Medical Guidelines for clinical practice for the evaluation and treatment of hypogonadism in adult male patients—2002 update. Endocr Pract 2002; 8:440-456. 3. Kaufman JM, Vermeulen A. The decline of androgen levels in elderly men and its clinical and therapeutic implications. Endocr Rev 2005;26:833-876. 4. Moskovic DJ, Araujo AB, Lipshultz LI, et al. The 20-year public health impact and direct cost of testosterone deficiency in U.S. men. J Sex Med 2013;10:562-569. 5. Seftel A. Male hypogonadism. Part II: etiology, pathophysiology, and diagnosis. Int J Impot Res 2006; 18:223-228. 6. Huhtaniemi I. Late-onset hypogonadism: current concepts and controversies of pathogenesis, diagnosis and treatment. Asian J Androl 2014;16:192-202. 7. Yang D, Dalton JE. A unified approach to measuring the effect size between two groups using SAS®. SAS Global Forum 2012:335-2012.
Conflicts of Interest: The authors report no conflicts of interest.
8. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011;46:399-424.
Funding: This study was funded by AbbVie, Inc. The sponsor was involved in the study design, data analysis, and interpretation of results and provided critical review of the report.
9. Bureau of Labor Statistics. Series CUUR0000SAM (medical care, all urban consumers). Available at: http://www.bls.gov/ cpi/home.htm. Accessed April 2015. J Sex Med 2017;14:88e97
95
Cost Burden of Primary or Secondary Hypogonadism 10. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130-1139. 11. Kaltenboeck A, Foster S, Ivanova J, et al. The direct and indirect costs among U.S. privately insured employees with hypogonadism. J Sex Med 2012;9:2438-2447. 12. Khandelwal N, Bhansali A, Johns B, et al. Assessment of healthcare utilization and costs in patients with primary and secondary hypogonadism. Poster PIH47 presented at the 21st Annual International Meeting of ISPOR. May 21-25, 2016; Washington, DC.
13. Haring R, Baumeister SE, Volzke H, et al. Prospective association of low serum total testosterone levels with health care utilization and costs in a population-based cohort of men. Int J Androl 2010;33:800-809. 14. Maggi M, Schulman C, Quinton R, et al. The burden of testosterone deficiency syndrome in adult men: economic and quality-of-life impact. J Sex Med 2007; 4:1056-1069. 15. Bhasin S, Cunningham GR, Hayes FJ, et al. Testosterone therapy in adult men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2006;91:1995-2010.
Appendix A. Hypogonadism, TRT, and testosterone panel codes* ICD-9-CM code Primary or secondary hypogonadism 253.2x 253.4x 253.7x 257.0x 257.1x 257.9x 608.3x 758.7x Age-related hypogonadism (excluded) 257.2x 302.71 302.72 799.81 TRT
Testosterone panel
Description Panhypopituitarism Other anterior pituitary disorders Iatrogenic pituitary disorders Testicular hyperfunction Postablative testicular hypofunction Unspecified testicular dysfunction Atrophy of testis Klinefelter syndrome Other testicular hypofunction Hypoactive sexual desire disorder With inhibited sexual excitement Decreased libido GPI codes 2310x, except 231000200003x, 23100030002900, 23100030302900, 23100020002900, 23100030102900, 23100005002900, 231000050001x, 23100020000105
CPT/HCPCS codes 11980, S0189, J3120, J3130, J3150, J3140, J1060, J1070, J1080, J0900
80414, 84403, 84402
CPT ¼ Current Procedural Terminology; GPI ¼ Generic Product Identifier; HCPCS ¼ Healthcare Common Procedure Coding System; ICD-9-CM ¼ International Classification of Diseases, Ninth Revision, Clinical Modification; TRT ¼ testosterone replacement therapy. *Selection of ICD-9-CM codes was based on a conservative assessment. The list of primary and secondary hypogonadism codes is not comprehensive and might not contain all ICD-9-CM codes associated with the condition but does include the most common codes used. A similar set of codes was used in a separate, recently published claims analysis.12 J Sex Med 2017;14:88e97
96
Grabner et al
Appendix B. Matching criteria used in study Variable Demographic Male Age ± 5 Index date year category Comorbidities and symptoms Quan-Charlson Comorbidity Index Hyperlipidemia or antihyperlipidemics Hypertension or antihypertensives
Coronary heart disease Diabetes or antidiabetics
Renal failure
Pain
Narcotic opioids Fractures
Fatigue Obstructive sleep apnea Insomnia or sleep disturbances or sleep medication
Obesity Hyperplasia of prostate Prostate cancer Depression or antidepressants Operational (known drivers of cost) Baseline number of office visits ± 2 Baseline number of other outpatient visits ± 2 Health plan type Region
Definition N/A Age at index date within 5 y Within 2-y blocks See Quan et al,10 2005 272.0x, 272.1x, 272.2x, 272.3x, 272.4x (ICD-9-CM); 39x (GPI) 401.0x, 401.1x, 401.9x, 402.00e405.99, 437.2x, 642.00e642.04, 642.10e 642.24, 642.70e642.94 (ICD-9-CM); 36x (GPI); G8506, 4010F, G8473, G8935, 4013F, 4480F (CPT/HCPCS) 410.xxe414.xx, 440.xx, 441.xx, 443.9x (ICD-9-CM) 250.xx (ICD-9-CM); V65.46, V58.67, V53.91, V45.85 (ICD-9-CM); 99.17 (ICD9 procedure); 27x (GPI); S5550eS5571, S9353, J1817, J1815, G9147, E0784, A9274 (CPT/HCPCS) 403.01, 403.11, 403.91, 404.02, 404.03, 404.12, 404.13, 404.92, 404.93, 585.3x, 585.4x, 585.5x, 585.6x, 585.9x, 586.xx, V42.0, V45.1x, V45.11, V45.12, V56.0x-V56.32, V56.8x (ICD-9-CM); 37x (GPI); S9361, 4221F (CPT/HCPCS) 720.xx, 721.2xe721.9x, 722.1xe722.3x, 722.5x, 722.6x, 722.70, 722.72, 722.73, 722.80, 722.82, 722.83, 722.90, 722.92, 722.93, 724.xx, 729.1x, 780.96, 338.xx, 307.80, 053.1x, 250.6x, 307.8x, 323.xx, 335.20, 335.34, 336.9x, 337.1x, 337.2x, 338.3x, 339.xx, 340.xx, 341.xx, 346.0xe346.9x, 350.xx, 351.xx, 353.xxe356.xx, 357.2x, 358.xx, 524.6x, 577.1x, 696.xx, 714.xx, 715.xx, 719.xx, 720.xx, 721.0xe722.4x, 722.71, 722.81, 722.91, 723.1x, 724.4x, 728.0x, 729.0x, 729.2x, 729.5x, 784.0x, 786.5x, 789.xx, 733.99, 733.14, 780.71, 820.8x, 820.9x, 951.4x, 952.xx, 953.4x, 955.5x e955.7x (ICD-9-CM); 88.81 (ICD-9 procedure); 22x, 64x, 65x, 66x, 67x, 7510x, 7520x (GPI); 62350, 62351, 62360e62362, 99601, 99602 96365 e96368, 96374e96379 (CPT/HCPCS) 65x (GPI) 800.xxe829.xx, 733.93e733.95, 733.8x, V15.51, V13.52, V54.0x, V54.1x, V54.8x, V54.9x, V66.4x, V67.4x, 813.41, 905.0xe905.7x (ICD-9-CM); 79.xx (ICD-9 procedure); L2132, L2102 -L2136, L2180eL2192, L3980 -L3999, 24500e24685, 28400 -28675 (CPT/HCPCS) 780.71, 780.70, 780.79, 300.5x (ICD-9-CM) 327.23 (ICD-9-CM) 780.51, 780.53, 780.57, 327.2x, 780.51, 780.52, 307.41, 307.42, 327.0x,780.53, 780.54, 307.43, 307.44, 327.1x, 327.20e327.23, 327.29, 327.3xe327.8x, 780.50, 780.55e780.59, 786.03 (ICD-9-CM); 602010x, 60100040x, 60100070x, 60200020x, 60204035x, 6025006000x, 60100025x, 60204080x,78512000000315, 6020407000x (GPI) 278.0x, 649.10e649.14, 793.91, V85.30eV85.4x, V85.54 (ICD-9-CM); 6125x (GPI) 600.xx (ICD-9-CM) 185.xx (ICD-9-CM) 300.4x, 301.12, 309.00, 309.1x, 311.xx (ICD-9-CM); 58x (GPI) N/A N/A N/A N/A
CPT ¼ Current Procedural Terminology; GPI ¼ Generic Product Identifier; HCPCS ¼ Healthcare Common Procedure Coding System; ICD-9-CM ¼ International Classification of Diseases, Ninth Revision, Clinical Modification; N/A ¼ not applicable.
J Sex Med 2017;14:88e97
97
Cost Burden of Primary or Secondary Hypogonadism
Appendix C. Baseline comorbidities and medications (within 12 months before index date) for matched patients with and without hypogonadism
Comorbidities, % Hypertension Hyperlipidemia Coronary heart disease Cerebrovascular disease Myocardial infarction Diabetes (mild to moderate) Diabetes (with complication) Renal failure Obesity Obstructive sleep apnea Asthma Chronic pulmonary disease Osteoarthritis Osteoporosis Depression Anxiety Fatigue HIV-AIDS Malignancy (except skin neoplasms) QCI, mean (SD) QCI, % 0e1 2 3 Baseline medications Antihypertensives Antihyperlipidemics Antidiabetics Mental health medications
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
46.3 53.6 7.3 2.0 0.7 16.2 1.6 6.8 3.9 3.9 10.5 6.7 10.7 2.4 19.1 21.3 25.5 0.9 1.6 0.5 (0.9)
46.3 53.6 7.3 1.7 0.8 15.7 1.5 6.8 3.9 3.9 9.0 7.4 10.7 1.2 19.1 19.4 25.5 0.8 1.6 0.5 (0.9)
69.9 21.1 9.0
70.9 20.7 8.4
33.0 35.9 13.1 33.6
30.1 31.4 11.6 29.0
P value
SMD
1.000 1.000 1.000 .408 .516 .493 .543 1.000 1.000 1.000 .008 .167 .952 <.0001 1.000 .012 1.000 .688 .825 .189 .580
0.00 0.00 0.00 1.54 1.21 1.28 1.13 0.00 0.00 0.00 4.96 2.57 0.11 9.34 0.00 4.69 0.00 0.75 0.41 0.00 2.27 1.02 2.21
.001 <.0001 .014 <.0001
6.33 9.46 4.58 9.98
QCI ¼ Quan-Charlson Comorbidity Index (modified from the Charlson Comorbidity Index for use in administrative databases)10; SMD ¼ standardized mean difference (in absolute value).
J Sex Med 2017;14:88e97
Clinical Characteristics, Health Care Utilization and Costs Among Men with Primary or Secondary Hypogonadism in a US Commercially Insured Population Michael Grabner, PhD,1 Amit Bodhani, MS,2 Nikhil Khandelwal, PhD,2 Swetha Palli, MS,3 Nicole Bonine, PhD,4 and Mohit Khera, MD5
ABSTRACT
Introduction: Hypogonadism is broadly associated with increases in chronic comorbid conditions and health care costs. Little is known about the specific impact of primary and secondary hypogonadism on health care costs. Aim: To characterize the health care cost and utilization burden of primary and secondary hypogonadism in a population of US men with commercial insurance. Methods: Newly diagnosed patients with International Classification of Diseases, Ninth Revision, Clinical Modification codes associated with specific medical conditions known to have a high prevalence of testosterone deficiency (ie, relating to primary or secondary hypogonadism) or who had fills for testosterone replacement therapy from January 1, 2007 through April 30, 2013 were identified in administrative claims data from the HealthCore Integrated Research Database. A cohort of patients without hypogonadism was matched on demographics and comorbidities. The matched hypogonadism and non-hypogonadism cohorts (n ¼ 5,777 in each cohort) were compared during a 12-month follow-up period. Main Outcome Measures: Direct health care expenditures and utilization were assessed for all causes and for hypogonadism-related claims. Costs included out-of-pocket patient expenditures and those paid by the insurer. Results: Hypogonadism and matched non-hypogonadism cohorts were similar in demographics (mean age ¼ 50 years) and diagnosed comorbid conditions in the 12 months preceding the index date. In the year after the index date, mean allcause expenditures for patients with hypogonadism increased by 62% (from $5,425 to $8,813) compared with 25% for the matched controls (from $4,786 to $5,992; P < .01 for follow-up difference between groups). Approximately 16% of total mean costs ($1,377), primarily outpatient and pharmacy costs, were identifiable as related to hypogonadism. Conclusion: These data from a population of US men with commercial insurance coverage showed a greater resource use burden for patients with primary and secondary hypogonadism compared with similar patients without hypogonadism. Additional management might be required to address unmet need and decrease the cost burden for patients with hypogonadism. J Sex Med 2017;14:88e97. Copyright 2016, The Authors. Published by Elsevier Inc. on behalf of the International Society for Sexual Medicine. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/). Key Words: Hypogonadism; Testosterone Deficiency; Retrospective Study; Health Care Utilization; Costs and Cost Analysis
INTRODUCTION Male hypogonadism has been defined as men with persistent testosterone deficiency and, according to Endocrine Received August 24, 2016. Accepted October 29, 2016. 1
HealthCore, Inc, Wilmington, DE, USA;
2
AbbVie Inc, Chicago, IL, USA;
3
CTI Clinical Trial and Consulting Services, Cincinnati, OH, USA;
4 5
Society guidelines, might be associated with symptoms such as decreased libido, sexual and erectile dysfunction, fatigue, lethargy, and mood disturbances.1,2 Decreased lean muscle
Copyright ª 2016, The Authors. Published by Elsevier Inc. on behalf of the International Society for Sexual Medicine. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/bync-nd/4.0/). http://dx.doi.org/10.1016/j.jsxm.2016.10.012
OptumRx, Irvine, CA, USA;
Baylor College of Medicine Medical Center, Houston, TX, USA
88
J Sex Med 2017;14:88e97
89
Cost Burden of Primary or Secondary Hypogonadism
mass and bone density are common and have been associated with fractures and frailty.2 Testosterone deficiency also has been associated with increased chronic comorbidities and health care costs.3,4 Specific causes of primary hypogonadism include genetic conditions (eg, Klinefelter syndrome), anatomic defects, infection, tumor, injury, iatrogenic causes (surgery or certain medications), and/or alcohol abuse. Secondary hypogonadism (low levels or inappropriately normal levels of follicle-stimulating hormone and luteinizing hormone in the presence of low testosterone) also can be caused by inherited conditions (eg, Kallmann syndrome), pituitary disorders, or hypothalamic lesions or disorders.5 The decrease of testosterone with increasing age is related to testicular and hypothalamic-pituitary function and is referred to as age-related hypogonadism or late-onset hypogonadism.1,6 There is limited published research estimating the specific impact of primary and secondary hypogonadism on health care costs. Understanding the relation of hypogonadism to costs over time requires isolating the impact in the context of other comorbidities. We attempted to advance this understanding by examining the costs of patients with primary and secondary hypogonadism longitudinally in a commercially insured population compared with matched controls.
AIMS The purpose of the present study was to estimate the direct cost burden of primary and secondary hypogonadism to health care payers and patients compared with a matched control group without hypogonadism.
METHODS We conducted a retrospective analysis using a US database of commercial health plan claims. Because the goal of the retrospective study was to include and summarize data for all eligible patients in the database, a predetermined sample size was not calculated. Throughout this study, researchers’ access was limited to data removed of all identifiers to ensure confidentiality. No study patients were contacted and informed consent was not obtained. An institutional review board did not review the study because only data in the format of a limited dataset were accessed and HealthCore maintains data use agreements with the covered entities in compliance with the Health Insurance Portability and Accountability Act of 1996.
Data Source All study data were derived from the HealthCore Integrated Research Database (HIRD), a repository of fully adjudicated medical and pharmacy claims on more than 37 million health plan members from across the United States. The commercial J Sex Med 2017;14:88e97
health plans included in the HIRD encompass health maintenance organizations, point of service, preferred provider organizations, indemnity plans, and Medicare Advantage plans. Dates queried spanned the period from January 1, 2006 through April 30, 2014. To allow for a 12-month pre-index period and a follow-up time of 12 months, the study intake period (containing the index date) was set at January 1, 2007 to April 30, 2013.
Patient Selection International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes and prescriptions for testosterone replacement therapy (TRT) were used to identify patients with hypogonadism. Adult men (18 years old) with commercial insurance coverage (employer-sponsored or Medicare Advantage) were included if they had at least two medical claims for conditions indicating hypogonadism (of any type; Appendix A presents a list of codes), or at least two pharmacy claims for TRT, or one claim for TRT and one claim for hypogonadism. The index date for patients with hypogonadism was defined as the date when patients first met at least one criterion for hypogonadism (from medical or pharmacy claims). For patients without hypogonadism, the index date was defined as the earliest date during a control patient’s enrollment period that created a successful match. The follow-up time was 12 months for all patients. Patients who died or lost coverage during the follow-up period were not included in the analysis because continuous enrollment was required for 12 months before and after the index date. A matched cohort of patients without hypogonadism, otherwise similar in demographic and clinical characteristics, was identified for comparison. A 1:1 matching algorithm was constructed to incorporate demographic characteristics, selected comorbidities, and baseline level of resource use (Appendix B presents a list of all matching criteria). The comorbidities for matching were frequently observed in the hypogonadism population and/or considered important in this setting based on a priori clinical input. Final criteria were selected with the goal of obtaining a close match on key confounding variables while preserving a large portion of the initial hypogonadism cohort. After creation of the matched cohorts, all patients with claims whose ICD-9-CM codes suggested the presence of age-related hypogonadism were excluded before the final analysis.
Analyses Descriptive statistics included means (SDs) for continuous data and absolute and relative frequencies for categorical data. For comparisons between the hypogonadism and nonhypogonadism groups, we used c2 tests for categorical variables, t-tests for normally distributed continuous variables, and Wilcoxon rank-sum tests for non-normally distributed continuous variables. All tests were two-sided using an a value of
90
Grabner et al
Figure 1. Patient attrition.
0.05. No adjustments for multiple comparisons were performed. Because the large samples available in administrative databases can lead to relatively trivial differences in magnitude being characterized as highly significant statistically, standardized mean differences (SMDs) were calculated between the cohorts to assist in interpretation of effect sizes.7 SMDs in absolute value of more than 10 were considered to have potential clinical significance.8 Analyses were performed using Excel 2007 (Microsoft, Redmond, WA, USA) and SAS 9.4 (SAS Institute, Cary, NC, USA).
MAIN OUTCOME MEASURES Resource utilization (hospital and emergency department admissions, outpatient visits including office visits and laboratory tests,
and pharmacy fills) and costs were assessed from medical and pharmacy claims during the 12-month baseline period and the 12month follow-up period. Costs were adjusted to 2014 US dollars using the medical care component of the Consumer Price Index.9 All-cause health care costs were reported in aggregate and segmented across places of service and included plan-paid costs and patient out-of-pocket expenditures. Resource utilization and costs were considered related to hypogonadism if the associated claims contained an ICD-9-CM diagnosis code indicative of hypogonadism or a procedure code for a testosterone panel or were pharmacy claims for TRT. All diagnosis fields, at any time during a hospital stay, were used to categorize inpatient hospitalizations. By definition, the non-hypogonadism cohort had no hypogonadism-related costs. Costs and utilization for the 12-month period after hypogonadism diagnosis were the main outcome metrics. J Sex Med 2017;14:88e97
91
Cost Burden of Primary or Secondary Hypogonadism
Table 1. Baseline demographics of study population*
Age (y), mean (SD)† Age (y), % 18e44 45e64 65 Type of health plan, %† HMO PPO Other Region, %† Northeast Midwest South West Index year, %† 2007e2008 2009e2010 2011e2012 2013 (JanuaryeApril) Treating MD specialty, % Primary care Urology Endocrinology Other or missing‡
Hypogonadism cohort before matching (n ¼ 9,463)
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
51 (11.6)
50 (11.1)
49 (11.1)
28.5 63.2 8.3
29.8 64.1 6.1
31.2 62.5 6.3
26.0 66.6 7.5
25.2 68.1 6.8
25.2 68.1 6.8
15.0 34.4 22.9 27.7
14.1 35.2 22.2 28.6
14.1 35.2 22.2 28.6
29.0 31.9 35.3 3.8
29.8 31.7 34.9 3.5
29.8 31.7 34.9 3.5
61.1 9.6 5.7 23.6
61.9 9.5 5.5 23.1
N/A N/A N/A N/A
HMO ¼ health maintenance organization; N/A ¼ not applicable; PPO ¼ preferred provider organization. *Categories might not total to 100% because of rounding. No statistically significant differences were noted between matched groups (P > .05) for demographic variables, and all absolute standardized mean differences were less than 10. No significance testing was conducted between the full hypogonadism cohort before matching and the matched groups. † Indicates a patient characteristic used in matching patients with and without hypogonadism (Appendix A). ‡ Includes other specialties (eg, geriatrics) and non-physicians.
RESULTS A total of 9,463 patients with claims for primary or secondary hypogonadism or TRT and who fulfilled all inclusion and exclusion criteria were included in the analysis (Figure 1). The mean age at index date was 51 years; 61% patients were treated in a primary care setting; and the largest geographic representation was in the Midwest (Table 1). Most patients received coverage under preferred provider organization health plans. After matching the patients with hypogonadism to a control cohort of patients without hypogonadism, 5,777 patients remained in each cohort. Overall, demographic characteristics in the full hypogonadism-TRT cohort were similar to those in the subset of matched patients with hypogonadism. The matched cohort was healthier than the full hypogonadism-TRT group, with fewer comorbid medical conditions (Table 2, Appendix C) and a lower mean Quan-Charlson comorbidity index score. As intended, the matched hypogonadism and nonhypogonadism cohorts were similar in patient characteristics. Hypertension and hyperlipidemia were common (46.3% and J Sex Med 2017;14:88e97
53.6%, respectively), and approximately 18% of patients had diabetes. Differences were seen in several variables not included in the list of matching requirements; for instance, baseline medication use, including mental health medication, was more common in the hypogonadism group than in the matched non-hypogonadism control group. However, in all cases, the absolute SMDs were below the threshold of 10. Meaningful clinical differences between the groups were not apparent at baseline for any of the demographic or clinical variables evaluated (Table 2). Health care utilization for patients with and without hypogonadism, while well matched at baseline, diverged in the year after the index date (Table 3). Patients with hypogonadism had a mean of 5.8 physician office visits and 8.4 visits to other outpatient services in the year after their diagnosis compared with 4.3 and 6.1, respectively, for patients without hypogonadism (P < .0001; SMD ¼ 28.8 and 22.8). Slightly more patients with than without hypogonadism were admitted to the hospital (7.9% vs 6.6%), although the magnitude of the difference was not large enough to be considered clinically relevant (SMD < 10). Patients with hypogonadism had an average of 28.2 pharmacy fills (12% of
92
Grabner et al
Table 2. Baseline comorbidities and medications (within 12 months before index date)
Comorbidities, % Hypertension* Hyperlipidemia* Coronary heart disease* Cerebrovascular disease Myocardial infarction Diabetes (mild to moderate)* Diabetes (with complication) Renal failure* Obesity* Obstructive sleep apnea* Asthma† Chronic pulmonary disease Osteoarthritis Osteoporosis† Depression* Anxiety† Fatigue* HIV-AIDS Malignancy (except skin neoplasms) QCI, mean (SD)* QCI, % 0e1 2 3 Baseline medications Antihypertensives† Antihyperlipidemics† Antidiabetics† Mental health medications†
Hypogonadism cohort before matching (n ¼ 9,463)
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
43.3 47.4 12.4 4.0 1.6 19.7 3.4 12.4 7.2 7.5 12.7 9.6 14.3 3.5 24.8 26.9 30.9 2.0 5.5 0.9 (1.6)
46.3 53.6 7.3 2.0 0.7 16.2 1.6 6.8 3.9 3.9 10.5 6.7 10.7 2.4 19.1 21.3 25.5 0.9 1.6 0.5 (0.9)
46.3 53.6 7.3 1.7 0.8 15.7 1.5 6.8 3.9 3.9 9.0 7.4 10.7 1.2 19.1 19.4 25.5 0.8 1.6 0.5 (0.9)
60.7 20.5 18.7
69.9 21.1 9.0
70.9 20.7 8.4
35.1 37.7 15.9 40.2
33.0 35.9 13.1 33.6
30.1 31.4 11.6 29.0
QCI ¼ Quan-Charlson Comorbidity Index (modified from Charlson Comorbidity Index for use in administrative databases).10 *Indicates a patient characteristic used in matching patients with and without hypogonadism (Appendix A). † P < .05 between matched groups, with absolute standardized mean differences less than 10 in all cases. No significance testing was conducted between the full hypogonadism cohort before matching and the matched groups. See Appendix C for individual P values and standardized mean differences.
which were related to hypogonadism) compared with 17.5 for patients without hypogonadism (P < .0001; SMD ¼ 46.4). In the year after hypogonadism diagnosis, mean all-cause expenditures for patients with hypogonadism increased by 62% (from $5,425 to $8,813) compared with 25% for matched controls (from $4,786 to $5,992); costs in the hypogonadism cohort were significantly higher compared with those in the matched control cohort ($8,813 vs $5,992; P < .001; SMD ¼ 15.0; Figure 2). Outpatient visits and pharmacy costs were the largest contributors to total cost and to cost increases. Approximately 15% of total costs, primarily outpatient and pharmacy costs, were identifiable as related to hypogonadism (mean ¼ $1,377; median ¼ $718 per patient per year; costs were considered related to hypogonadism if the associated claims contained an ICD-9-CM diagnosis code indicative of hypogonadism or a procedure code for a testosterone panel or were pharmacy claims for TRT).
DISCUSSION This retrospective study from a commercial US health plan describes a population of more than 5,000 men with newly diagnosed primary or secondary hypogonadism. We found that patients with primary and secondary hypogonadism had increased health care utilization and cost during the year after their diagnosis compared with matched controls without diagnosed hypogonadism but with a similar baseline comorbidity burden. These results are consistent with those of a previous study using retrospective data, also in the United States, but from a different source.11 Kaltenboeck et al11 examined a prevalent, rather than incident, hypogonadism cohort (including primary, secondary, and age-related etiologies) and used risk adjustment rather than direct matching to estimate control group costs. Annual risk-adjusted all-cause direct costs in that study were J Sex Med 2017;14:88e97
93
Cost Burden of Primary or Secondary Hypogonadism
Table 3. Health care utilization during 12 months after index date
Patients with 1 inpatient hospitalization, n (%)* With 1 hypogonadism-related hospitalization, % Patients with 1 emergency room visit, n (%) With 1 hypogonadism-related visit, % Physician office visits per patient per year, mean (SD)†,‡ Total visits related to hypogonadism, % Other outpatient visits per patient per year, mean (SD)†,‡ Total visits related to hypogonadism, % Pharmacy, number of fills per patient per year, mean (SD)†,‡ Total fills for hypogonadism, %
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
457 8.3 750 1.6 5.8 12.2 10.4 14.6 24.8 12.4
384 N/A 713 N/A 4.3 N/A 6.2 N/A 17.0 N/A
(7.9) (13.0) (5.8) (11.6) (22.7)
(6.6) (12.3) (5.0) (9.9) (20.2)
N/A ¼ not applicable. *Any emergency room visit resulting in hospital admission was counted toward inpatient hospitalizations. † Absolute standardized mean difference greater than 10. ‡ P < .0001 between groups.
$9,291 for hypogonadism (any type) vs $5,248 for nonhypogonadism controls. Of note, because the study by Kaltenboeck et al focused on employed men, the investigators could estimate indirect costs from disability and absenteeism, which added an additional 22% to the risk-adjusted cost difference. Comorbidity profiles for patients with hypogonadism (any type)
Figure 2. All-cause health care costs for 12 months before and after index date (mean; 2014 US dollars). Costs during 1-year follow-up (POST) were significantly higher in the hypogonadism cohort than in the non-hypogonadism cohort ($8,813 vs $5,992; P < .001; standardized mean difference ¼ 15.0). J Sex Med 2017;14:88e97
were comparable to those seen in the present study. A second, more recent, study using administrative claims data came to a similar conclusion regarding cost differences between patients with vs without primary or secondary hypogonadism; in a sample of 7,386 matched patients from January 2006 through June 2014, all-cause total costs ($13,686 vs $7,551) and medical costs ($10,675 vs $5,769) were significantly higher in patients with hypogonadism.12 Other past studies, using different data sources and methods, also have found an increase in utilization and costs for patients with hypogonadism.13,14 Strengths of the present study include the large number of analyzed patients and the creation of a well-defined comparison cohort by matching. Because the intake period began in 2007, all patients began hypogonadism treatment after the 2006 Endocrine Society guidelines became available.15 Tight matching on several patient characteristics was possible given the large initial sample and the detail captured in the administrative claims. The small proportion of patients at least 65 years old in our analysis reflects well on our efforts to focus on a primary and secondary hypogonadism population; nevertheless, external validation of the set of ICD-9-CM codes that was used would be valuable. The matching algorithm was successful in creating cohorts of primary and secondary hypogonadism and non-hypogonadism that were similar at baseline in all demographic and clinical aspects. We incorporated baseline resource utilization and other operational variables (plan type and region) known to affect health care costs in the matching algorithm. The overall pre-match hypogonadism-TRT population was less healthy compared with the more strictly defined matched group, suggesting that resource utilization and costs could be even higher in the overall hypogonadism population compared with the non-hypogonadism population. Limitations of the study are primarily a result of reliance on medical claims data, which lack clinical context and are
94
subject to potential coding, billing, and recording errors and other inconsistencies. The lack of availability of certain clinical data meant we could not replicate current diagnostic criteria for hypogonadism. Specifically, testosterone and gonadotropin levels were not used to determine inclusion in the present analysis because laboratory data were available only for a small subset of patients in this data source. Clinical symptoms reported by patients that did not lead to a billable health care encounter would have been missed. Therefore, the control cohort could include a small percentage of men with low testosterone levels who might not have been diagnosed with hypogonadism. The two cohorts also included patients based on the presence of TRT claims only, without requiring the presence of ICD-9-CM codes for primary or secondary hypogonadism; however, those with ICD-9-CM codes for agerelated hypogonadism were excluded (a small proportion of patients might have had age-related hypogonadism without the presence of a related code). Similarly, important risk factors such as obesity are typically under-reported in claims data. Conversely, the presence of a diagnosis code on a medical claim does not guarantee the presence of a disease, because the diagnosis code might be incorrectly coded or included as a rule-out criterion. We required at least two claims for hypogonadism or TRT on distinct dates to limit this bias. Claims databases also only include patients with specific health care coverage, in this case private commercial health plans. Notably, patients without health insurance or those who are covered under public programs (such as Medicare or Medicaid) are not part of the analyzed population. Therefore, the ability to generalize these results to the entire US population might be limited. Commercially insured individuals are more likely to be currently employed and tend to be healthier than the general population, especially if they continue working. Further research is necessary to determine how primary and secondary hypogonadism contributes to increased health care costs over time. The constellation of symptoms observed in patients with hypogonadism—decreased energy, loss of muscle, depressed mood, fractures, and frailty—can lead to physical inactivity and aggravate the attendant risk of diabetes and cardiovascular disease. This study was limited to 1 year of follow-up after hypogonadism diagnosis, so a full picture of the economic burden of hypogonadism might require longer observation. More research also is needed to investigate the potential benefit of TRT in mitigating the burden of hypogonadism. Correspondence Author: Michael Grabner, PhD, HealthCore, Inc, 123 Justison Street, Suite 200, Wilmington, DE 19801, USA. Tel: 530-400-5978; E-mail: [email protected]
Grabner et al
STATEMENT OF AUTHORSHIP Category 1 (a) Conception and Design Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera (b) Acquisition of Data Michael Grabner; Swetha Palli; Nicole Bonine (c) Analysis and Interpretation of Data Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera Category 2 (a) Drafting the Article Michael Grabner; Swetha Palli; Nicole Bonine (b) Revising It for Intellectual Content Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera Category 3 (a) Final Approval of the Completed Article Michael Grabner; Amit Bodhani; Nikhil Khandelwal; Swetha Palli; Nicole Bonine; Mohit Khera
REFERENCES 1. Bhasin S, Cunningham GR, Hayes FJ, et al. Testosterone therapy in men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2010;95:2536-2559. 2. Petak SM, Nankin HR, Spark RF, et al. American Association of Clinical Endocrinologists Medical Guidelines for clinical practice for the evaluation and treatment of hypogonadism in adult male patients—2002 update. Endocr Pract 2002; 8:440-456. 3. Kaufman JM, Vermeulen A. The decline of androgen levels in elderly men and its clinical and therapeutic implications. Endocr Rev 2005;26:833-876. 4. Moskovic DJ, Araujo AB, Lipshultz LI, et al. The 20-year public health impact and direct cost of testosterone deficiency in U.S. men. J Sex Med 2013;10:562-569. 5. Seftel A. Male hypogonadism. Part II: etiology, pathophysiology, and diagnosis. Int J Impot Res 2006; 18:223-228. 6. Huhtaniemi I. Late-onset hypogonadism: current concepts and controversies of pathogenesis, diagnosis and treatment. Asian J Androl 2014;16:192-202. 7. Yang D, Dalton JE. A unified approach to measuring the effect size between two groups using SAS®. SAS Global Forum 2012:335-2012.
Conflicts of Interest: The authors report no conflicts of interest.
8. Austin PC. An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 2011;46:399-424.
Funding: This study was funded by AbbVie, Inc. The sponsor was involved in the study design, data analysis, and interpretation of results and provided critical review of the report.
9. Bureau of Labor Statistics. Series CUUR0000SAM (medical care, all urban consumers). Available at: http://www.bls.gov/ cpi/home.htm. Accessed April 2015. J Sex Med 2017;14:88e97
95
Cost Burden of Primary or Secondary Hypogonadism 10. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005;43:1130-1139. 11. Kaltenboeck A, Foster S, Ivanova J, et al. The direct and indirect costs among U.S. privately insured employees with hypogonadism. J Sex Med 2012;9:2438-2447. 12. Khandelwal N, Bhansali A, Johns B, et al. Assessment of healthcare utilization and costs in patients with primary and secondary hypogonadism. Poster PIH47 presented at the 21st Annual International Meeting of ISPOR. May 21-25, 2016; Washington, DC.
13. Haring R, Baumeister SE, Volzke H, et al. Prospective association of low serum total testosterone levels with health care utilization and costs in a population-based cohort of men. Int J Androl 2010;33:800-809. 14. Maggi M, Schulman C, Quinton R, et al. The burden of testosterone deficiency syndrome in adult men: economic and quality-of-life impact. J Sex Med 2007; 4:1056-1069. 15. Bhasin S, Cunningham GR, Hayes FJ, et al. Testosterone therapy in adult men with androgen deficiency syndromes: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab 2006;91:1995-2010.
Appendix A. Hypogonadism, TRT, and testosterone panel codes* ICD-9-CM code Primary or secondary hypogonadism 253.2x 253.4x 253.7x 257.0x 257.1x 257.9x 608.3x 758.7x Age-related hypogonadism (excluded) 257.2x 302.71 302.72 799.81 TRT
Testosterone panel
Description Panhypopituitarism Other anterior pituitary disorders Iatrogenic pituitary disorders Testicular hyperfunction Postablative testicular hypofunction Unspecified testicular dysfunction Atrophy of testis Klinefelter syndrome Other testicular hypofunction Hypoactive sexual desire disorder With inhibited sexual excitement Decreased libido GPI codes 2310x, except 231000200003x, 23100030002900, 23100030302900, 23100020002900, 23100030102900, 23100005002900, 231000050001x, 23100020000105
CPT/HCPCS codes 11980, S0189, J3120, J3130, J3150, J3140, J1060, J1070, J1080, J0900
80414, 84403, 84402
CPT ¼ Current Procedural Terminology; GPI ¼ Generic Product Identifier; HCPCS ¼ Healthcare Common Procedure Coding System; ICD-9-CM ¼ International Classification of Diseases, Ninth Revision, Clinical Modification; TRT ¼ testosterone replacement therapy. *Selection of ICD-9-CM codes was based on a conservative assessment. The list of primary and secondary hypogonadism codes is not comprehensive and might not contain all ICD-9-CM codes associated with the condition but does include the most common codes used. A similar set of codes was used in a separate, recently published claims analysis.12 J Sex Med 2017;14:88e97
96
Grabner et al
Appendix B. Matching criteria used in study Variable Demographic Male Age ± 5 Index date year category Comorbidities and symptoms Quan-Charlson Comorbidity Index Hyperlipidemia or antihyperlipidemics Hypertension or antihypertensives
Coronary heart disease Diabetes or antidiabetics
Renal failure
Pain
Narcotic opioids Fractures
Fatigue Obstructive sleep apnea Insomnia or sleep disturbances or sleep medication
Obesity Hyperplasia of prostate Prostate cancer Depression or antidepressants Operational (known drivers of cost) Baseline number of office visits ± 2 Baseline number of other outpatient visits ± 2 Health plan type Region
Definition N/A Age at index date within 5 y Within 2-y blocks See Quan et al,10 2005 272.0x, 272.1x, 272.2x, 272.3x, 272.4x (ICD-9-CM); 39x (GPI) 401.0x, 401.1x, 401.9x, 402.00e405.99, 437.2x, 642.00e642.04, 642.10e 642.24, 642.70e642.94 (ICD-9-CM); 36x (GPI); G8506, 4010F, G8473, G8935, 4013F, 4480F (CPT/HCPCS) 410.xxe414.xx, 440.xx, 441.xx, 443.9x (ICD-9-CM) 250.xx (ICD-9-CM); V65.46, V58.67, V53.91, V45.85 (ICD-9-CM); 99.17 (ICD9 procedure); 27x (GPI); S5550eS5571, S9353, J1817, J1815, G9147, E0784, A9274 (CPT/HCPCS) 403.01, 403.11, 403.91, 404.02, 404.03, 404.12, 404.13, 404.92, 404.93, 585.3x, 585.4x, 585.5x, 585.6x, 585.9x, 586.xx, V42.0, V45.1x, V45.11, V45.12, V56.0x-V56.32, V56.8x (ICD-9-CM); 37x (GPI); S9361, 4221F (CPT/HCPCS) 720.xx, 721.2xe721.9x, 722.1xe722.3x, 722.5x, 722.6x, 722.70, 722.72, 722.73, 722.80, 722.82, 722.83, 722.90, 722.92, 722.93, 724.xx, 729.1x, 780.96, 338.xx, 307.80, 053.1x, 250.6x, 307.8x, 323.xx, 335.20, 335.34, 336.9x, 337.1x, 337.2x, 338.3x, 339.xx, 340.xx, 341.xx, 346.0xe346.9x, 350.xx, 351.xx, 353.xxe356.xx, 357.2x, 358.xx, 524.6x, 577.1x, 696.xx, 714.xx, 715.xx, 719.xx, 720.xx, 721.0xe722.4x, 722.71, 722.81, 722.91, 723.1x, 724.4x, 728.0x, 729.0x, 729.2x, 729.5x, 784.0x, 786.5x, 789.xx, 733.99, 733.14, 780.71, 820.8x, 820.9x, 951.4x, 952.xx, 953.4x, 955.5x e955.7x (ICD-9-CM); 88.81 (ICD-9 procedure); 22x, 64x, 65x, 66x, 67x, 7510x, 7520x (GPI); 62350, 62351, 62360e62362, 99601, 99602 96365 e96368, 96374e96379 (CPT/HCPCS) 65x (GPI) 800.xxe829.xx, 733.93e733.95, 733.8x, V15.51, V13.52, V54.0x, V54.1x, V54.8x, V54.9x, V66.4x, V67.4x, 813.41, 905.0xe905.7x (ICD-9-CM); 79.xx (ICD-9 procedure); L2132, L2102 -L2136, L2180eL2192, L3980 -L3999, 24500e24685, 28400 -28675 (CPT/HCPCS) 780.71, 780.70, 780.79, 300.5x (ICD-9-CM) 327.23 (ICD-9-CM) 780.51, 780.53, 780.57, 327.2x, 780.51, 780.52, 307.41, 307.42, 327.0x,780.53, 780.54, 307.43, 307.44, 327.1x, 327.20e327.23, 327.29, 327.3xe327.8x, 780.50, 780.55e780.59, 786.03 (ICD-9-CM); 602010x, 60100040x, 60100070x, 60200020x, 60204035x, 6025006000x, 60100025x, 60204080x,78512000000315, 6020407000x (GPI) 278.0x, 649.10e649.14, 793.91, V85.30eV85.4x, V85.54 (ICD-9-CM); 6125x (GPI) 600.xx (ICD-9-CM) 185.xx (ICD-9-CM) 300.4x, 301.12, 309.00, 309.1x, 311.xx (ICD-9-CM); 58x (GPI) N/A N/A N/A N/A
CPT ¼ Current Procedural Terminology; GPI ¼ Generic Product Identifier; HCPCS ¼ Healthcare Common Procedure Coding System; ICD-9-CM ¼ International Classification of Diseases, Ninth Revision, Clinical Modification; N/A ¼ not applicable.
J Sex Med 2017;14:88e97
97
Cost Burden of Primary or Secondary Hypogonadism
Appendix C. Baseline comorbidities and medications (within 12 months before index date) for matched patients with and without hypogonadism
Comorbidities, % Hypertension Hyperlipidemia Coronary heart disease Cerebrovascular disease Myocardial infarction Diabetes (mild to moderate) Diabetes (with complication) Renal failure Obesity Obstructive sleep apnea Asthma Chronic pulmonary disease Osteoarthritis Osteoporosis Depression Anxiety Fatigue HIV-AIDS Malignancy (except skin neoplasms) QCI, mean (SD) QCI, % 0e1 2 3 Baseline medications Antihypertensives Antihyperlipidemics Antidiabetics Mental health medications
Matched hypogonadism cohort (n ¼ 5,777)
Matched non-hypogonadism cohort (n ¼ 5,777)
46.3 53.6 7.3 2.0 0.7 16.2 1.6 6.8 3.9 3.9 10.5 6.7 10.7 2.4 19.1 21.3 25.5 0.9 1.6 0.5 (0.9)
46.3 53.6 7.3 1.7 0.8 15.7 1.5 6.8 3.9 3.9 9.0 7.4 10.7 1.2 19.1 19.4 25.5 0.8 1.6 0.5 (0.9)
69.9 21.1 9.0
70.9 20.7 8.4
33.0 35.9 13.1 33.6
30.1 31.4 11.6 29.0
P value
SMD
1.000 1.000 1.000 .408 .516 .493 .543 1.000 1.000 1.000 .008 .167 .952 <.0001 1.000 .012 1.000 .688 .825 .189 .580
0.00 0.00 0.00 1.54 1.21 1.28 1.13 0.00 0.00 0.00 4.96 2.57 0.11 9.34 0.00 4.69 0.00 0.75 0.41 0.00 2.27 1.02 2.21
.001 <.0001 .014 <.0001
6.33 9.46 4.58 9.98
QCI ¼ Quan-Charlson Comorbidity Index (modified from the Charlson Comorbidity Index for use in administrative databases)10; SMD ¼ standardized mean difference (in absolute value).
J Sex Med 2017;14:88e97