The effect of age and sex on cost of inpatient facility encounters among patients with multiple sclerosis

Abstract

Objective:

To explore the effect of age and sex on cost of all-cause and multiple sclerosis (MS)-related inpatient facility encounters.

Methods:

Adult patients with an initial MS diagnosis were identified from a national managed-care database (IMS LifeLink Health Plans Database). The analysis included newly diagnosed MS patients with 12 months insurance eligibility before and after their first MS diagnosis. Inpatient facility encounters (stays) were analyzed for all-cause and MS-related events (ICD-9-CM = 340.XX), other demyelinating CNS disease (ICD-9-CM = 341.XX), rehabilitation (ICD-9-CM = V57.89), and a group of symptom-related diagnoses. Costs and length of stay were evaluated using a general linear model controlling for age and sex.

Results:

A total of 57,236 patients met study criteria; 74.3% were female. Mean age for females was 45.5 years and for males it was 47.5 years. In total, 17.0% had an all-cause inpatient stay in the 360-day post index, and 3.2% had an in patient stay with a MS relapse-related diagnosis as primary discharge diagnosis. Additional MS-related diagnoses that led to inpatient stays included other demyelinating CNS disease (0.3%), symptom-related diagnoses (1.0%), and rehabilitation (1.1%). All-cause inpatient cost was higher for males vs females across all age groups; however, cost for females increased at a greater rate as age increased (p = 0.0007). Symptom-related inpatient cost was flat for males, was lower for females than males at an average age of 30, and was greater for females than males at an average age of 60 (p = 0.0199). Cost for MS inpatient stays ($11,931), other demyelinating CNS-related stays ($14,931), and rehabilitation ($23,643) did not differ by age and sex. The average cost for any MS-related relapse inpatient stay was $13,761 and varied with increasing age (p < 0.0001).

Conclusion:

Burden of illness for relapse among MS patients is substantial. Costs vary by age and sex depending on the discharge diagnosis. Inclusion of symptom-related and rehabilitation inpatient stays may account for an under-recognized proportion of total expenditures.

Keywords::

• Multiple sclerosis
• Disease modifying drug (DMD)
• Relapse
• Acute care
• Inpatient stays
• Costs

Introduction

Multiple sclerosis (MS) is an autoimmune disorder in which the fatty myelin sheaths around the axons of the nerves of the central nervous system (CNS) are demyelinated, leading to a spectrum of symptoms that are often disabling, costly, and unpredictable1. Prevalence in the US is ∼400,000 (almost 2.5 million worldwide) with ∼200 new cases diagnosed each week2. MS is usually diagnosed between the ages of 20–40, which means it can affect the lives of people over many years as life expectancy of MS patients is similar to that in the general population.

In recent years researchers have noted a disparity in MS incidence, relapse rates, and disease characteristics between females and males. Results of a 2006 study conducted in Canada showed that the female-to-male MS incidence ratio by year of birth had grown to exceed 3.2:1 over a 50 year period3. According to Dunn and Steinman4, ‘the worrisome increase in the incidence of MS in women is the subject of intense scrutiny’ (p. 635). A recent study designed to evaluate the effect of sex differences in the percentages of patients experiencing MS relapses reported that relapse frequency was 17.7% higher in females vs males and concluded that females are pre-disposed to higher relapse activity than males5. Bove et al.6 investigated the influence of age and sex on clinical and radiologic outcomes among a MS cohort. Patients were grouped to represent years before (38–46 years) and after (54–62 years) age 50. While there were changes with male and older cohorts showing worse disease severity, brain atrophy, and more rapid progression, there was ‘no interaction between age and sex suggestive of an effect of reproductive aging on clinical or radiologic progression’ (p. 1). The influence of sex and age on cost of inpatient encounters, a component of MS relapse, has not been well studied.

MS-related relapses are a focus of attention because relapse is associated with significant cost. Ivanova et al.7 reported on cost of relapse among employees with MS in a US population. Results of the 2-year study showed that employees who were adherent to disease modifying drugs (DMDs) had a lower rate of severe MS relapses vs non-adherent employees (12.5% vs 19.5%; p = 0.0127). Risk-adjusted total costs excluding DMD costs were significantly lower among adherent employees compared with non-adherent employees ($14,095 vs $16,638; p = 0.0475)7. These studies evaluated average costs of relapse over time; there are few published evaluations on the cost of relapse at an encounter level.

Burden of illness for MS relapse is substantial. The cost of inpatient facility encounters (stays) can result in significant insurer and patient responsibility. The objective of this study was to explore the effect of age and sex on cost of all-cause and multiple sclerosis (MS)-related inpatient facility encounters.

Patients and methods

Study design

This was a retrospective claims analysis for an MS patient cohort derived from a national claims database.

Data source

Patients with commercial insurance coverage were selected from the IMS LifeLink Health Plans Database. LifeLink is an anonymous, patient-centric, national managed care database representing more than 71 million people with employer-based health insurance. In addition to all medical and pharmacy claims data for study patients, LifeLink also includes demographic variables (age, sex, region of the US), eligibility by month, and the adjudicated (i.e., final after any adjustments) payment for services.

Patient selection

Patients were required to have an ‘initial’ MS diagnosis (ICD9-CM: 340.xx) defined as the first MS diagnosis in one of up to four diagnoses fields in the database; with no MS diagnosis for at least 12 months prior. The date of the first MS diagnosis was the study index date. Patients were required to have 12 months of continuous eligibility (insurance coverage) before and after the index date. Index dates occurred between January 1, 2006 and July 30, 2012. Patients were also required to be at least 18 years old and less than or equal to 64 years of age on the index date.

Record selection

For inpatient stays, the diagnosis on the last facility record was taken to be the discharge diagnosis (as per data vendor recommendations at the time of the data extraction) and was used for record selection. This was assumed to be the primary reason for the admission. Inpatient stays that began and ended during the 12-month post-index period were included.

Inpatient stays were categorized by type of diagnosis; MS diagnosis (ICD9-CM = 340.xx), a diagnosis for ‘other demyelinating CNS disease’ (ICD9-CM = 341.xx), or a symptom-related diagnosis that was identified based on review of published literature (unpublished). The review used ‘multiple sclerosis’ and ‘symptoms’ as search terms. The potential MS symptom-related conditions identified by the review are presented in Table 1. The symptom-related codes were evaluated as a group and labeled as ‘symptom-related’. In addition, a preliminary analysis of diagnosis code frequencies actually appearing in the data led to the inclusion of the diagnosis code for ‘care involving other specified rehabilitation procedure; multiple training or therapy’ (ICD9-CM = V57.89).

Table 1. Codes used for symptom indicators that are potentially related to MS.

Symptom	Code	Description
Urinary incontinence	788.3X	Urinary Incontinence
Malaise and fatigue	780.7X	Malaise and Fatigue
Balance disorders	780.4X	Dizziness and Giddiness Light-headedness; Vertigo NOS
Altered mental state	780.97	Altered Mental Status Change in mental status
Numbness	782.0	Disturbance of Skin Sensation Anesthesia of skin; Burning or prickling sensation; Hyperesthesia; Hypoesthesia; Numbness; Paresthesia; Tingling
Muscle spasm	728.85	Spasm of Muscle
Stiffness	719.50	Stiffness of Joint, NEC
Depression	296.2X	Major Depressive Disorder, Single Episode {Depressive psychosis} {Endogenous depression} {Involutional melancholia} {Manic-depressive psychosis or reaction, depressed type} {Monopolar depression} {Psychotic depression} single episode or unspecified
	296.3X	Major Depressive Disorder, Recurrent Episode Any condition classifiable to 296.2, stated to be recurrent
	300.4X	Dysthymic Disorder Anxiety depression; Depression with anxiety; Depressive reaction; Neurotic depressive state; Reactive depression
	311.XX	Depressive Disorder, NEC Depressive disorder NOS; Depressive state NOS; Depression NOS
Constipation	564.0X	Constipation
Optical neuritis	377.3X	Optic Neuritis
Rehabilitation	V57.89	Care involving other specified rehabilitation procedure; Multiple training or therapy

Outcome measurement

The focus of the analysis was on the inpatient facility encounter (stay). Inpatient records were identified based on the data provider’s confinement identifier. The confinement identifier uses the first facility room and board record through the last service date billed for the same facility and assigns all records with a date of service between these dates as part of the inpatient event. The measures of interest were the cost per person per stay and the length of stay per person per stay. The question addressed was whether cost or length of stay per stay varied depending on age and/or sex. The claim ‘allowed’ amount was used for all cost calculations. The allowed amount is the total amount paid including any insured’s liability (e.g., co-pay, deductible, and co-insurance) for a stay. Inpatient costs per stay were summed for the following record types: Management Services (M; provider costs related to the direct evaluation or management of a patient while under inpatient care); Surgical Services (S); Facility Services (F; room and board cost of an institutional provider); Ancillary Services (A; cost of incidental services for direct patient care (e.g., X-ray, lab, etc.)). Costs were adjusted to December 2012 using the monthly Medical Services component for all urban consumers of the Consumer Price Index, published by the US Bureau of Labor Statistics8.

Analysis

Categorical variables were summarized using frequencies and percentages. Continuous variables were summarized using means (with confidence intervals), standard deviations and medians. Patients were evaluated in the 12-month post-index period for inpatient stays for the following discharge diagnoses: All-cause, MS diagnosis, Other demyelinating diagnoses, MS symptom-related diagnoses, Rehabilitation diagnoses, Any potential MS relapse (included MS diagnosis, other demyelinating diagnoses, symptom-related diagnoses and rehabilitation).

A general linear model with robust variance estimation was used to derive estimated costs for inpatient stays controlling for age and sex. The model used a Gamma distribution with a log link to address potential issues with non-normal distributions for costs and length of stay. The same model was used for all evaluations (dependent measure = age, sex, and an age by sex interaction). Estimated costs were derived using the parameters of the estimating equation for males and females at ages 30, 40, 50, and 60. When sex and the sex by age interaction were not significant, cost was estimated at the proportions of males and females in the sample. When age and the age by sex interaction were not significant, age was estimated at the mean in the sample being evaluated and sex was estimated at the proportion of males and females in the sample. An overall model and models for each diagnosis type were evaluated independently controlling for age and sex. All analyses were conducted using SAS version 9.4 for Windows (SAS Institute Inc., Cary, NC).

Results

A total of 64,022 patients had 12 months of pre- and post-eligibility around their first MS diagnosis date. Of these, 6786 (3.8%) were removed for not meeting the age requirement, leaving 57,236 available for analysis. Of the total, 43,064 (75.2%) were female and 14,172 (24.8%) were male. The mean age for females was 45.4 (SD = 10.4) years and 46.0 (SD = 11.0) years for males (p < 0.0001).

Table 2 presents the descriptive analysis for age, sex, the percentage of subjects with a stay, and the mean cost per stay (overall and by diagnosis type) for the 12-month post-index period. Overall, 17.0% had an all-cause inpatient stay in the 360-day post-index period, and 3.2% had an inpatient stay with a MS diagnosis (ICD-9-CM = 340.XX) as the primary discharge diagnosis. Other MS-related diagnoses that led to inpatient stays included other demyelinating central nervous system (CNS) disease (0.3%), symptom-related diagnoses (1.0%), and rehabilitation (1.1%). When all potentially MS-related diagnoses are combined, any potential MS relapse-related diagnosis was experienced by 5.2% of the cohort.

Table 2. Medical price index adjusted (December 2012) inpatient cost per inpatient stay per person with age and sex descriptives (n = 57,236).

Category	With stays		Mean cost with CI			Median	Age and sex descriptives
	n	%	Mean (SD)	Lower CI	Upper CI		Category	%	Mean (SD)
All cause	9728	17.0	17,283 (21,340)	16,859	17,707	11,289	Males	25.7%	47.4 (11.5)
							Females	74.3%	45.5 (10.9)
MS	1828	3.2	11,939 (11,570)	11,409	12,470	9692	Males	26.3%	42.2 (11.3)
							Females	73.7%	42.4 (10.7)
Other demyelinating disease	186	0.3	15,058 (13,855)	13,054	17,063	11,252	Males	36.6%	43.1 (11.9)
							Females	63.4%	43.0 (12.1)
Symptom-related	579	1.0	9878 (16,966)	8493	11,263	6967	Males	22.1%	44.6 (13.1)
							Females	77.9%	44.3 (10.8)
Rehabilitation	628	1.1	23,694 (27,269)	21,557	25,830	15,901	Males	34.2%	49.4 (11.0)
							Females	65.8%	49.3 (10.4)
Any potential MS relapse-related diagnosis	2990	5.2	13,760 (16,567)	13,166	14,354	9889	Males	27.5%	44.4 (11.9)
							Females	72.5%	44.0 (11.0)

Of the reasons for inpatient stays, the average cost was highest for rehabilitation, while the lowest cost was for symptom-related conditions. There was considerable variation in the means, as shown by the wide confidence intervals. Median cost was always lower than mean cost. The percentage of males was highest for the other demyelinating disease and rehabilitation diagnoses categories. Age was greatest within the all-cause and rehabilitation diagnoses categories.

Model results are presented in Table 3 and Figures 1 and 2. Table 3 presents a summary of the models by presenting the probabilities for significant effects for the age, sex and age by sex interactions. If the age by sex effect is significant it means that interpretation of the cost of inpatient stays for that category depends on both age and sex (i.e., cost differ by both age and sex). If none of the effects are significant then the costs are not affected by age, sex, or the age by sex interaction, and the mean for the sample is the best estimate of the cost for that category.

Figure 1. Cost and length of stay per inpatient event for all-cause inpatient stays (n = 9725; 17.0%).

Figure 2. Cost and length of stay per inpatient event for symptom-related inpatient stays (n = 579; 1.0%).

Table 3. Table of p-values* for significant effects.

Diagnosis type	Cost models			LOS models
	Age	Sex	Age*Sex	Age	Sex	Age*Sex
All cause (n = 9725; Figure 1)	<0.0001	<0.0001	0.0007	<0.0001	<0.0001	0.0010
MS (n = 1828; No Figure)	NS	NS	NS	<0.0001	0.7109	0.5747
Other demyelinating disease (n = 186; No Figure)	NS	NS	NS	0.2501	0.0080	0.0094
Any symptom (n = 579; Figure 2)	0.9944	0.0322	0.0199	0.0133	0.0918	0.0824
Rehabilitation (n = 628; No Figure)	NS	NS	NS	NS	NS	NS
Any potential MS relapse-related diagnosis (n = 2990)	<0.0001	0.5671	0.7933	<0.0001	0.2586	0.9461

*p-values were derived from a general linear model using a gamma distribution and a log link.

The model for all-cause inpatient cost controlling for age, sex, and the age by sex interaction showed a significant age by sex interaction (p ≤ 0.0007; Figure 1). Males had greater estimated costs than females, with costs increasing by age for both male and female patients; however, cost for females increased at a faster rate. There was a greater difference in estimated cost at age 30 (Males = $17,493 vs Females = $13,265) compared to age 60 (Males = $21,721 vs Females = $19,565). The observed changes appear to be consistent with the observed lengths of stay, supporting the association between length of stay and costs that conventional wisdom would suggest.

The model for inpatient symptom-related stays showed an even more distinctive age by sex interaction (p = 0.0199; Figure 2). Estimated costs for males were almost unchanged, regardless of age (range = $9456–$9469), while cost for females ranged from $7658 at 30 years of age to $12,896 at 60 years of age. Supplementary Tables A1 and A2 present the estimated means and confidence intervals for each age and gender combination. These tables provide greater detail regarding differences between estimated costs at the selected ages.

Age and sex were not significant predictors in the inpatient stay models for MS (estimated cost $11,931), other demyelinating disease (estimated cost $14,931), or rehabilitation (estimated cost 23,643). Costs associated with rehabilitation inpatient stays were highest among the categories. The average cost per rehabilitation was $23,643 and the average LOS was 16.7 days. The ‘any MS-related’ model showed only a significant (p < 0.0001) age effect (i.e., sex and age by sex effects were not significant) with relapse cost ranging from $12,037 at 30 years of age, $13,172 at 40 years of age, $14,415 at 50 years of age, and $15,775 at 60 years of age. Numbers may differ slightly from the unadjusted values due to rounding of the estimates obtained from the general linear model.

The relationship for length of inpatient stay (LOS) varied relative to the cost models. The statistical tests for LOS for all-cause inpatient stays showed a similar pattern for the cost model, with a significant age by sex interaction (p = 0.0010). For all-cause stays, males had a greater length of stay at all ages; however, females LOS increased at a greater rate as age increased (Figure 1). Similar LOS and the same directional effect was observed for other demyelinating disease (p = 0.0094 for the age by sex interaction). For other demyelinating disease the estimated LOS between 30–60 years of age ranged from 5.4–8.3 for females and declined for males from 7.7 to 6.0 (data not shown). Age and sex were not significant modifiers for rehabilitations stays (mean estimated LOS = 16.7 days). For the other three categories, MS (mean LOS by year at 30 years of age = 4.7 days; 40 years of age = 5.8 days; 50 years of age = 6.4; 60 years of age = 7.5), symptom (Figure 2), and any MS-related (mean LOS by year at 30 years of age = 5.8 days; 40 years of age = 7.0 days; 50 years of age = 8.6; 60 years of age = 10.5), only age was associated with LOS.

Discussion

This study assessed all-cause and MS-related inpatient facility encounter length of stay and cost, evaluating differences based on patient age and sex in the 12 months after an initial diagnosis for MS. Comparison of cost from the current analysis to previous studies is difficult due to differences in sample composition, study design, study dates, and cost reporting methods. The closest comparison is a study by O’Brien et al.9. The O’Brien et al. study used all-payer data from five states and estimated the average cost of a MS-related (ICD9-CM = 340.XX) hospital inpatient stay at $8782 in 2002 US$ and a rehabilitation facility stay at $14,943. When these values are adjusted to 2012 dollars using the Medical Services component of the consumer price index, a value of $13,257 is obtained for inpatient costs and $22,557 for rehabilitation. The adjusted inpatient estimate is lower than the all-cause estimate of $17,283 found in this analysis. However, differences could reflect costs being averaged over different types of facilities. It is also possible that there are greater costs of technology given the length of time since O’Brien et al.’s analysis was conducted. Interestingly, the adjusted rehabilitation cost ($22,557) is more similar to the finding in this analysis ($23,694). It is noteworthy that the average age and percentage of females in both analyses were similar. O’Brien et al. also report that the cost of managing a relapse is more than 6-times higher when patients receive inpatient care as part of their treatment, reinforcing the importance of understanding the cost of inpatient encounters.

Modeling the influence of age and sex on inpatient costs provided interesting results. The relationship between cost of inpatient stays, with age and sex, varied by the type of diagnosis. Analysis of the cost of inpatient stays for all-cause and symptom-related diagnoses showed a significant age by sex effect. The all-cause LOS analysis showed the same trend as the cost analysis. The all-cause age and sex-adjusted LOS is also directionally consistent with the length of stay from the Centers for Disease Control and Prevention CDC/NCHS National Hospital Discharge Survey10. The CDC data show, for all US discharges (all diagnoses excluding newborns), that males have a longer length of stay (LOS) than females at all ages. Female length of stay begins to approach the LOS for males between 45–64 years of age. Males at 45–64 years of age have an average LOS of 5.2 days and females 4.8 days. While the direction of effects was similar, on average this analysis showed a LOS that was greater by ∼1–2 days for all sub-groups. This may not be unexpected given that this is a cohort of MS diagnosed patients that may have greater overall severity then the general population. The impression left by this analysis is that the results for all-cause stays is not unexpected.

What might be somewhat less expected was the pattern that was observed for symptom-related inpatient stays. Symptom-related inpatient stays stand out for two reasons. First, the cost for male inpatient stays remains relatively flat as age increases in contrast to the typical pattern for male stays in the all-cause group. Second, the female group exceeded the male group costs at approximately age 40 and was over $3000 greater by age 60. The cost model was predictive using age and sex and the LOS model was predictive using age. It is also noteworthy that these symptom-related stays are not typically included when considering the costs of MS-related inpatient facility encounters. Further exploration of inpatient stays that may be associated with MS progression is warranted and might account for an under-recognized portion of MS-related costs.

An examination of data shows that the highest cost was for rehabilitation. Rehabilitation inpatient stays may be lengthy, which has a significant effect on the cost of the stay. In this study there was no significant difference in rehabilitation utilization based on age, sex, or age and sex. A noteworthy observation from this analysis is that rehabilitation inpatient stays represent another area that could easily be overlooked in analysis of MS-related costs or relapse rate analyses. Inpatient events for these types of diagnoses were sufficiently common to warrant consideration as potential markers for MS relapse. Comprehensive inpatient rehabilitation has been shown to be beneficial for maintaining functional capacity and reducing loss of function for patients who have had a disabling exacerbation or deterioration11^,12.

Cost of inpatient stays with an MS diagnosis, other demyelinating disease, and rehabilitation stays did not appear to be affected by age and sex. This lack of effect by age and sex suggests similarity of treatment when these more specific diagnoses are used. It appears likely that, as the specificity of the reason for a stay increases, the homogeneity of the group also increases, which in turn reduces the differences observed between sexes. It is noteworthy, however, that, except for rehabilitation, there is a significant effect of age on LOS, which may be in part due to the slightly greater mean age in the rehabilitation group. So, while LOS increased with age, there was not a corresponding increase in costs. Analysis of hospital charge-master data to further elucidate differences between MS and other demyelinating disease inpatient stays by age group appears warranted.

Limitations

This analysis used a non-experimental retrospective design and, therefore, may not be generalizable. Presence of a symptom, rehabilitation, or other demyelinating diagnosis code does not assure that it was related to MS nor that it was an MS relapse. The claims data used in this study were collected for administrative purposes, not specifically for research studies. Identification of inpatient claims was done based on indicators provided by the data vendor. This grouping may include a range of facilities, resulting in a limitation of the coding for facilities. Consequently, there may be coding errors and missing data which limit inferences. Given the consequences of falsely reporting procedures and cost, and the size of the available sample, these errors are expected to be limited. The inclusion criteria requiring hospitalizations to begin and end within the calendar year of observation reduced the likelihood that long-term hospitalizations were included in the analysis. Inclusion of these hospitalizations may have led to greater average costs. Other variables that may have a significant effect on costs were not included in the model (e.g., region, payer type, duration of disease, marital status). The regional distribution of the overall sample included 31.0% from the East, 32.5% Midwest, 22.4% South, and 14.1% from the West. Results could differ by region and analysis of a data-set with a different mix of patients could result in different results. Additional analysis of variables not included in this analysis is warranted.

Conclusions

Costs for inpatient visits among MS patients were substantial. Differences in inpatient costs varied by age and sex depending on the discharge diagnosis. Inclusion of rehabilitation and symptom-related data may account for a previously under-recognized cost of relapse. Further examination of the cost of care for MS patients should include these cost categories.

Transparency

Declaration of funding

Funding was provided by EMD Serono, Inc., One Technology Place, Rockland, MA 02370 and Pfizer, Inc. 235 East 42nd Street, NY, NY 10017.

Declaration of financial/other relationships

DMM is a former employee of EMD Serono, Inc. and ALP is an employee of EMD Serono, Inc. CMK received research support from EMD Serono, Inc and Pfizer, Inc.

Acknowledgments

Employees of EMD Serono, Inc. participated in the design of the study, the analysis of the data, and the preparation of the paper. Chris Kozma, Ph.D., participated in the design of the study, conducted the analysis and participated in preparation of the paper. Mike Dickson, PhD, and Lavonda Miley, PhD, independent consultants, participated in the preparation of the manuscript.