HomeLarge Type Edition
HOME ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
PubMed
Right arrow PubMed Citation
The Journals of Gerontology Series A: Biological Sciences and Medical Sciences 61:411-415 (2006)
© 2006 The Gerontological Society of America

Medication Utilization in an Urban Homebound Population

Ian M. Kronish, Alex D. Federman, R. Sean Morrison and Jeremy Boal

1 Division of General Internal Medicine and 2 Brookdale Department of Geriatrics and Adult Development, Mount Sinai School of Medicine, New York, New York.

Address correspondence to Ian M. Kronish, MD, Box 1087, One Gustave L. Levy Place, New York, NY 10029. E-mail: ian.kronish{at}mssm.edu


   Abstract
Top
 Abstract
Methods
 Results
 Discussion
 References
 
Background. The number of medically homebound adults has grown with the aging of the U.S. population, yet little is known about their health care utilization. We sought to characterize the health status and medication utilization of an urban cohort of homebound adults and to identify factors associated with medication use in this population.

Methods. We performed a retrospective cross-sectional analysis of 415 patients enrolled in a primary care program for homebound adults in New York City during October 2002. Numbers of medications were obtained from formularies corroborated by home visits. For patients without prescription insurance, medication out-of-pocket costs were estimated according to average wholesale pricing. Sociodemographic and disease characteristics were obtained by chart abstraction.

Results. The median age was 83 years (range 25–106 years). Seventy-seven percent of patients were female, 63% were non-white, and 28% spoke Spanish. Sixty-four percent of patients had Medicaid. The cohort had a mean of 8.2 (range 1–27, standard deviation 4.5) medications prescribed per month. Multivariate analysis showed that increasing age was associated with fewer medications (p <.001). Charlson comorbidity score was positively associated with number of medications (p <.001), whereas Activities of Daily Living score, a measure of functional dependence, was not. Twenty-seven percent of the cohort lacked prescription drug coverage. The total number of medications per month among the uninsured patients was 7.4 (standard deviation 4.4). Estimated median monthly out-of-pocket cost for the uninsured patients was $223 (range $1–$1512).

Conclusions. For homebound patients without prescription drug coverage, medication use may represent substantial financial burden. Additional research is needed to determine whether out-of-pocket medication costs represent a barrier to care in this population.

ELDERLY and disabled patients constitute a growing segment of the U.S. population (1). Nursing homes have absorbed some of the burden of caring for these patients (2–4). Yet, many chronically ill patients prefer to remain in their homes (5,6). By the late 1990s, approximately two million adults, half of whom were 65 years or older, were estimated to be permanently homebound (7). Like other disabled patients, homebound patients are expected to utilize a disproportionate amount of health care resources. As a group, however, their utilization of resources has not yet been well-studied.

In recent years, the fastest growing portion of out-of-pocket health care costs has been prescription medications. Between 1992 and 2000, the average price per prescription among older persons rose 48%. By 2000, drug expenses consumed 14% of the average Social Security benefit, up from 8% in 1992 (8). Prescription drugs represent the highest out-of-pocket spending category among patients with chronic illnesses (9). The growing burden of prescription drugs is due not only to price increases, but to an increasing volume of medications prescribed (10). A number of studies have examined prescription utilization within the Medicare (11–13) and the nursing home population (14). Less is known about the burden of prescriptions among medically homebound adults (15,16).

In this study, we examined medication use in a cohort of homebound adults. We also assessed the association between medication use and the socioeconomic and disease characteristics of this population. We hypothesized that increased age and increased burden of comorbid disease would be associated with increased medication use. Further, we tested whether functional dependence, a distinctive feature of this population, was associated with medication use. Although out-of-pocket medication costs were not measured directly, we provided an estimate of medication costs borne by patients who lacked prescription insurance. We hypothesized that uninsured homebound patients faced a high cost burden from prescribed medications.


   METHODS
Top
 Abstract
 Methods
Results
 Discussion
 References
 
Patient Enrollment and Setting
We conducted a retrospective, cross-sectional analysis of data on the 417 patients enrolled in the Mount Sinai Hospital Visiting Doctors Program (VDP) as of October 31, 2002. The median length of time in the program was 19 months. The VDP is an urban primary care program for permanently homebound adults in which patients receive home visits by either a physician or nurse practitioner for routine outpatient care. Additional home-based services include phlebotomy, physical therapy, subspecialty consultations, and access to a social worker with expertise in issues pertaining to homebound patients such as procurement of home services. Patients, however, do not receive any direct prescription medication benefits from the program.

Patients are eligible for the program if they are expected to permanently meet the Medicare definition of homebound and if they reside in the New York City borough of Manhattan. Medicare defines homebound as an inability to leave home without considerable effort, including dependence on assistive devices, special transportation, or other persons (17). Patients are referred to the program from community-based organizations, private physicians, and home health agencies. Fewer than one percent of patients in the VDP are ever discharged from the program. The most common reason for referral was dementia (29%). Other reasons included refusal to go the doctor or nonadherence with prior care (16%), frailty (9%), end-stage neuromuscular disease (7%), severe arthritis (6%), stroke (5%), end-stage cardiopulmonary disease (5%), and morbid obesity (4%). The study was approved by the Mount Sinai School of Medicine Institutional Review Board.

Data Collection
Data for the study were collected by abstraction of VDP charts. The VDP physicians and nurse practitioners complete a standardized intake form for all new patients, by interviewing either patients or their surrogates. Demographic information collected includes gender, age, ZIP code, language, race/ethnicity, health insurance and prescription drug coverage, and whether the patient lives alone. Patients with 24-hour home attendants are classified as not living alone. Prescription drug coverage data were missing for nine patients. It is standard protocol for VDP providers to keep a record of prescription insurance status in the patient's medical chart. Census data from 2000 were used to identify the median household income within each patient's ZIP code.

Providers also collect information pertaining to disease burden, including chronic illnesses and limitations of Activities of Daily Living (ADL). ADL scores are a widely used measure of functional status (18). Patients are scored from 0 to 2 for dependence in each of 8 ADLs (bathing, dressing, feeding, toileting, transferring, continence, walking, and grooming). A score of 0 indicates no assistance is required, 1 indicates partial dependence, and 2 indicates full dependence. For our analyses, we calculated each patient's ADL score and Charlson comorbidity score (19). The Charlson score is a validated measure of risk for death in patients with chronic illness. In the original validation study, the percentage of patients who died within 1 year according to the Charlson score were: "0", 8%; "1", 25%; "2", 48%; and "3", 59%.

Medication use was derived from a database of patient medication lists which was updated during each physician home visit, and thus offered a cross-section of monthly prescription use. Providers review all medication bottles during each home visit before updating the database. In New York State, Medicaid and various insurance companies cover the expense of some over-the-counter medications (including stool softeners and vitamins) if these are prescribed by a physician. Accordingly, no distinction was made between over-the-counter and prescription-only medications. A therapeutic drug class was assigned to each medication according to the classification scheme of the American Hospital Formulary Service (20).

Data Analysis
We summed the number of medications for each patient and estimated the monthly costs of medications according to the average wholesale price as listed in the 2002 Drug Topics Red Book (21). Generic costs were utilized if a drug had a generic version available. Costs for medications prescribed as needed were estimated by halving the maximum prescribed amount.

Eight medications were excluded from the analysis because either their prices were not listed in the Red Book or the clinical database from which the patients' medications were queried included inadequate descriptions of their dosages or vehicles to accurately estimate pricing. Medications excluded were the following: lidocaine/prilocaine, methylcellulose, oxygen, clindamycin 2%, miconazole, clotrimazole, tioconazole, phenobarbital/atropine. Using a local pharmacy's price estimates for average usages of these medications, we found that none of the excluded medications would have affected the total monthly costs of any individual patient by more than 3%. Two patients were excluded from the study because the dosages of their opioid prescriptions were unknown. Unlike the eight medications excluded for lacking pricing in the Red Book with little impact on overall monthly costs, these medications could have altered total monthly costs of individual patients by more than two standard deviations depending on the total doses consumed. After these corrections, the data set comprised 415 patients using a total of 3409 medications.

Univariate analysis was performed using the chi-square test for proportions and the pooled t test and one-way analysis of variance (ANOVA) for continuous variables. Correlations between two continuous, non-normally distributed groups were measured using the Spearman rank method. Linear regression was used to assess the multivariable association between number of medications used and patients' demographic and health status characteristics. Variables having p <.05 for their association with number of medications used on univariate analysis were included in the regression model. Race/ethnicity was also included to test whether there was any bias in prescribing pattern based on minority status. Prescription insurance status but not health insurance status was entered as there is significant overlap between these variables. Further, ADL score was forced into the model because functional dependence is a distinctive feature of a homebound population. All statistical analyses were performed using Stata 6.0 statistical software (Stata Corporation, College Station, TX).


   RESULTS
Top
 Abstract
 Methods
 Results
Discussion
 References
 
Patient Characteristics
Socioeconomic characteristics of the cohort are summarized in Table 1. The median age was 83 years (range 25–106 years). Thirteen percent of the patients were younger than 65 years, and another 13% were older than 95 years. Three quarters were women, 37% were white, 28% were African American, and 31% were Latino. Nearly one third of the cohort was Spanish speaking (28%). Fourteen percent of the group resided within a ZIP code where the median household income was less than $20,000. Nearly one third of the patients lived alone (30%).


View this table:
[in this window]
[in a new window]
  		Table 1. Characteristics of Homebound Patients and Associated Medication Use.

 
All patients had medical insurance. Nine percent had Medicare alone, and 52% had combined Medicare and Medicaid coverage. Seventy-three percent of patients had prescription drug coverage, mostly in the form of Medicaid. Only 9% had prescription insurance other than Medicaid. These other forms of insurance included commercial insurance and Elderly Pharmaceutical Insurance Coverage (EPIC), a New York State–funded prescription medication plan for low-income New York Medicare beneficiaries who do not qualify for Medicaid.

The most common medical comorbidities included hypertension, dementia, osteoarthritis, depression, and diabetes (Table 2). The most commonly clustered comorbidities were hypertension and dementia (25%), hypertension and osteoarthritis (25%), dementia and osteoarthritis (21%), and osteoarthritis and depression (16%). Patients had a mean ADL score of 8.7 (standard deviation 5.4), indicating marked functional dependence. Eight percent had the minimum total ADL score of 0, and 17% had a maximum total ADL score of 16. The median number of comorbid diseases per patient was 4 (range 0–13). Fifty percent of patients had a Charlson score ≥2, indicating a very elevated risk of mortality within 1 year.


View this table:
[in this window]
[in a new window]
  		Table 2. Frequency of Most Common Medical Comorbidities.

 
Prescribed Medication Utilization
The median number of medications prescribed for each patient each month was 8 (range 1–27). Medication utilization by drug therapeutic class is summarized in Table 3. The most commonly prescribed therapeutic classes for which drugs were prescribed included antihypertensives, vitamins, laxatives, aspirin, and antidepressants. As vitamins, laxatives, and minerals do not always qualify as prescribed medications in other states, we performed our analysis again without these medications included. The median number of medications per patient without the aforementioned drugs included was six medications per patient each month. Patients with dementia were prescribed 1.2 fewer medications per month as compared to patients without dementia (p =.007).


View this table:
[in this window]
[in a new window]
  		Table 3. Medication Utilization by Most Frequent Therapeutic Drug Class.

 
Estimated Drug Expenditures for Patients Without Prescription Drug Coverage
The homebound adults in our sample who lacked prescription drug coverage used a median of 7 (range 1–19) medications per month. On the basis of the average wholesale price, we estimated their median out-of-pocket monthly prescription drug costs to be $233 (interquartile range $126–$353). The full range of their out-of-pocket costs was from $1 to $1512 per month.

Multivariate Analysis
In a linear regression model of number of medications used among the total study population, we found lower age, female gender, and higher Charlson score to be associated with greater use of medications (Table 4). There was no significant association between number of medications used and total ADL score. Spearman correlations were also performed to examine the relationship between age and measures of disease burden. Unlike in the general population, ADL score (p =.4) and Charlson score (p =.3) were not correlated with age in this homebound cohort.


View this table:
[in this window]
[in a new window]
  		Table 4. Multivariate Linear Regression Analysis of the Association Between Medication Use and Homebound Patients' Demographic and Health Status Characteristics (N = 349).

 

   DISCUSSION
Top
 Abstract
 Methods
 Results
 Discussion
References
 
This study provides a detailed description of the disease and medication utilization characteristics of an urban medically homebound population. We document high levels of medication utilization by homebound patients in this program. The level of medication use of homebound patients (median of eight per month) may in fact exceed that of patients in long-term residential care facilities, which ranges from five to eight medications (22,23). Furthermore, the level of medication use of homebound patients is greater than that of a more general group of Medicare patients receiving home health care services who took a median of five medications per month (24).

We found that the number of medications used by homebound patients declined as age increased, an association that contradicts the pattern observed among adults in community-based epidemiological studies (25). The explanation for the inverse association in the homebound group is likely multifactorial. First, among the homebound patients, burden of disease as measured by Charlson score or ADL score was evenly distributed among different age groups (instead of increasing with age), indicating that young homebound patients were as burdened by poor health as were older ones. Further, practitioners for the homebound may have been more sensitive to the risks of polypharmacy in the elderly patients, which may have served to moderate their prescribing patterns in older age groups. Finally, prior research of medication use in a nursing home population has shown that there is an association between increasing cognitive impairment, increasing age, and decreasing numbers of medications (26,27). This association was confirmed in our homebound group.

Even though disability and loss of functional independence are distinctive features of homebound persons, ADL score was not related to increased number of medications in this group. Of the two measures of disease burden, only Charlson score had a significant association. Functional dependence does, however, lead to higher costs for home services such as home attendants and visiting nurses, costs that are borne variably by patients and their families (28).

More than one quarter of our study population lacked prescription drug coverage. For these patients, the median monthly cost of $223 represented an approximation of out-of-pocket drug costs. For comparison, median out-of-pocket monthly prescription costs among Medicare patients with four or more chronic diseases (based on data from the Medicare Current Beneficiary Survey) has been estimated at $29 in a 1997 study (29).

Our study has some notable limitations. First, the study was restricted to a single program for permanently homebound patients residing in New York City. Our cohort had a high percentage of lower income patients, and medications were prescribed by physicians and nurse practitioners who were specialized in treating patients in their homes. Thus, disease burden, drug utilization, and insurance patterns may not be representative of other homebound patients. Nevertheless, our study population was very diverse and included a broad range of income levels, ethnicities, and age groups. The most common comorbidities were typical of other reported homebound populations (30). Providers may have been better trained to restrict their prescribing practices out of concern for the risks of polypharmacy in the predominantly geriatric population. Hence, this study may actually underestimate the overall volume and burden of medication use in this population.

Second, costs were calculated according to average wholesale price. In many cases, average wholesale price does not represent the actual price to the consumer (31). For instance, managed care prescription plans can frequently obtain discounts of up to 20% off the average wholesale price (32). Pharmacies who sell direct to consumers, however, may sell medications at more than the average wholesale price. Moreover, in this study, generic prices were utilized whenever possible as the study database did not distinguish between generic and brand name prescriptions. Because uninsured patients were likely to have bought more expensive brand-name prescriptions in some instances and were not receiving discounted prices through managed care or Veterans Administration plans, we may have underestimated actual out-of-pocket costs to these patients.

Finally, we were not able to account for all possible measures of health status that might be important predictors of medication utilization. For example, we were not able to include a measure of symptom burden. Further, we relied on ADL score as our measure of functional status and did not include measures of physical performance such as gait speed. However, by including ADL score and Charlson index, we were able to test for two important measures.

A significant proportion of homebound adults lacked insurance to cover their medications' costs. Previous studies have shown that the financial cost of caring for homebound persons reverberates through family support networks. For example, among caregivers of seriously ill patients, up to one third will spend their family's savings on medical care and one fifth will be forced to quit their job or make another major life change (33). Patients have also been documented to self-restrict medications when cost is an issue (34–37). At the same time, physicians are not well-trained in identifying cost issues (38,39). Accordingly, future studies should examine how homebound patients and their physicians cope with the high costs of medications (40). Studies should also assess whether physicians are trained to accommodate costs of care and patient goals of care in the way they implement practice guidelines, especially with regard to prescribing for homebound patients with severe functional limitations and high medication costs. Homebound patients may benefit from specialized practitioners who are attuned to the risks of polypharmacy and to the burden of out-of-pocket costs of medications. Although the Medicare prescription drug benefit enacted by Congress in 2003 may partially ease the burden of paying for prescriptions for seniors in this population, special attention will need to be paid to younger homebound patients who are ineligible for Medicare.


   Acknowledgments
Top
 Abstract
 Methods
 Results
 Discussion
 References
 
We thank Katherine Ornstein, MPH, and Albert Siu, MD, MSPH, for their helpful comments on the manuscript.


   Footnotes
Top
 Abstract
 Methods
 Results
 Discussion
 References
 
Decision Editor: Darryl Wieland, PhD, MPH

Received July 21, 2005

Accepted December 8, 2005


   References
Top
 Abstract
 Methods
 Results
 Discussion
 References
 

  1. Health Care Financing Administration. A Profile of Medicare: Chartbook. Rockville, MD: Department of Health and Human Services; 1998:10.
  2. Spillman BC, Lubitz J. New estimates of lifetime nursing home use: have patterns of use changed? Med Care. 2002;40:965-975.[Medline]
  3. Lakdawalla D, Goldman DP, Bhattacharya JP, Hurd MD, Joyce GF, Panis CW. Forecasting the nursing home population. Med Care. 2003;41:8-20.[Medline]
  4. Besdine RW, Rubenstein LZ, Cassel C. Nursing home residents need physicians' services. Ann Intern Med. 1994;120:616-618.[Free Full Text]
  5. Cummings JE, Hughes SL, Weaver FM, et al. Cost-effectiveness of Veterans Administration hospital-based home care. A randomized clinical trial. Arch Intern Med. 1990;150:1274-1280.[Abstract/Free Full Text]
  6. Taler G. House calls for the 21st century. J Am Geriatr Soc. 1998;46:246-248.[Medline]
  7. Kassner E, Bectel RW., Mid-life and older Americans with disabilities: who gets help? A chartbook. Washington, DC: Public Policy Institute, American Association of Retired Persons; 1998.
  8. Davidoff F. The heartbreak of drug pricing. Ann Intern Med. 2001;134:1068-1071.[Free Full Text]
  9. Hwang W, Weller W, Ireys H, Anderson G. Out-of-pocket medical spending for care of chronic conditions. Health Aff. 2001;20:267-278.[Abstract/Free Full Text]
  10. Dubois RW, Chawla AJ, Neslusan CA, Smith MW, Wade S. Explaining drug spending trends: does perception match reality? Health Aff. 2000;19:231-239.[Medline]
  11. Davis M, Poisal J, Chulis G, Zarabozo C, Cooper B. Prescription drug coverage, utilization, and spending among Medicare beneficiaries. Health Aff. 1999;18:231-243.[Abstract]
  12. Soumerai SB, Ross-Degnan D. Inadequate prescription-drug coverage for Medicare enrollees—a call to action. N Engl J Med. 1999;340:722-728.[Free Full Text]
  13. Gross D, Brangan N. Medicare beneficiaries and prescription drug coverage: gaps and barriers. Washington, DC: Public Policy Institute, American Association of Retired Persons; 1999:1–10.
  14. O'Neill C, Groom L, Avery AJ, Boot D, Thornhill K. Explaining variations in prescribing costs: results from a comparison of nursing home patients with matched pairs living in the community. J Clin Pharm Ther. 1999;24:427-432.[Medline]
  15. Fried TR, Wachtel TJ, Tinetti ME. When the patient cannot come to the doctor: a medical housecalls program. J Am Geriatr Soc. 1998;46:226-231.[Medline]
  16. Fleming BB, Pulliam CC, Perfetto EM, Hanlon JT, Bowling JM. Medication use by home health patients. J Geriatric Drug Ther. 1993;7:33-45.
  17. Wachtel TJ, Gifford DR. Eligibility for home care certification: what clinicians should know. J Gen Intern Med. 1998;13:705-709.[Medline]
  18. Katz S, Ford AB, Moskowitz RW, et al. Studies of illness in the aged. The index of the ADL: a standardized measure of biological and psychosocial function. JAMA. 1963;185:914-919.[Abstract/Free Full Text]
  19. Charlson ME, Ponpei P, Ales KL, et al. A new method for classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373-383.[Medline]
  20. McEvoy GK, ed. American Hospital Formulary Service, 2005. Bethesda, MD: American Society of Health-System Pharmacists; 2005.
  21. Cardinale V, ed. 2002 Drug Topics Red Book. Montvale, NJ: Medical Economics Co.; 2002.
  22. Williams BR, Nichol MB, Lowe B, Yoon PS, McCombs JS, Margolies J. Medication use in residential care facilities for the elderly. Ann Pharmacother. 1999;33:149-155.[Abstract]
  23. Robers PA. Extent of medication use in U.S. long-term-care facilities. Am J Hosp Pharm. 1988;45:93-100.[Abstract]
  24. Meredith S, Feldman PH, Frey D, et al. Possible medication errors in home healthcare patients. J Am Geriatr Soc. 2001;49:719-724.[Medline]
  25. Mueller C, Schur S, O'Connell J. Prescription drug spending: the impact of age and chronic disease status. Am J Public Health. 1997;87:1626-1629.[Abstract/Free Full Text]
  26. Chiang L, Hirsch SH, Reuben DB. Predictors of medication prescription in nursing homes. J Am Med Dir Assoc. 2000;1:97-102.[Medline]
  27. Schmader KE, Hanlon JT, Fillenbaum GG, Huber M, Pieper C, Horner R. Medication use patterns among demented, cognitively impaired and cognitively intact community-dwelling elderly people. Age Ageing. 1998;27:493-501.[Abstract/Free Full Text]
  28. Fried TR, Bradley EH, Williams CS, Tinetti ME. Functional disability and health care expenditures for older persons. Arch Intern Med. 2001;161:2602-2607.[Abstract/Free Full Text]
  29. Sambamoorthi U, Shea D, Crystal S. Total and out-of-pocket expenditures for prescription drugs among older persons. Gerontologist. 2003;43:345-359.[Abstract/Free Full Text]
  30. Ganguli M, Fox A, Gilby J, Belle S. Characteristics of rural homebound older adults: a community-based study. J Am Geriatr Soc. 1996;44:363-370.[Medline]
  31. Gencarelli DM. Average wholesale price for prescription drugs: is there a more appropriate pricing mechanism? NHPF Issue Brief. 2002;775:1-19.
  32. Steinberg EP, Gutierrez B, Momani A, Boscarino JA, Neuman P, Deverka P. Beyond survey data: a claims-based analysis of drug use and spending by the elderly. Health Aff. 2000;19:198-211.[Abstract]
  33. Covinsky KE, Goldman L, Cook EF, et al. The impact of serious illness on patients' families. SUPPORT Investigators. Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatment. JAMA. 1994;272:1839-1844.[Abstract/Free Full Text]
  34. Kennedy J, Erb C. Prescription noncompliance due to cost among adults with disabilities in the United States. Am J Public Health. 2002;92:1120-1124.[Abstract/Free Full Text]
  35. Steinman MA, Sands LP, Covinsky KE. Self-restriction of medications due to cost in seniors without prescription coverage. J Gen Intern Med. 2001;16:793-799.[Medline]
  36. Stuart B, Grana J. Ability to pay and the decision to medicate. Med Care. 1998;36:202-211.[Medline]
  37. Sharkey JR, Ory MG, Browne BA. Determinants of self-management strategies to reduce out-of-pocket prescription medication expense in homebound older people. J Am Geriatr Soc. 2005;53:666-674.[Medline]
  38. Heisler M, Wagner TH, Piette JD. Clinician identification of chronically ill patients who have problems paying for prescription medications. Am J Med. 2004;116:753-758.[Medline]
  39. Reichert S, Simon T, Halm EA. Physicians' attitudes about prescribing and knowledge of the costs of common medications. Arch Intern Med. 2000;160:2799-2803.[Abstract/Free Full Text]
  40. Rector TS, Venus PJ. Do drug benefits help Medicare beneficiaries afford prescribed drugs? Health Aff. 2004;23:213-222.[Abstract/Free Full Text]




This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Services
Right arrow Download to citation manager
PubMed
Right arrow PubMed Citation

HOME ARCHIVE SEARCH TABLE OF CONTENTS