Skip to main content
  • Research article
  • Open access
  • Published:

“Implications of cost-sharing for observation care among Medicare beneficiaries: a pilot survey”



Medicare beneficiaries hospitalized under observation status have significant cost-sharing responsibilities under Medicare Part B. Prior work has demonstrated an association between increased cost-sharing and health care rationing among low-income Medicare beneficiaries. The objective of this study was to explore the potential impact of observation cost-sharing on future medical decision making of Medicare beneficiaries.


Single-center pilot cohort study. A convenience sample of Medicare beneficiaries hospitalized under observation status care was surveyed.


Out of 144 respondents, low-income beneficiaries were more likely to be concerned about the cost of their observation stay than higher-income respondents (70.7% vs29.3%, p = 0.015). If hospitalized under observation status again, there was a trend among low-income beneficiaries to request completion of their workup outside of the hospital (56.3% vs 43.8%), and to consider leaving against medical advice (AMA) (100% vs 0%), though these trends were not statistically significant (p = 0.30).


The results of this pilot study suggest that low-income Medicare beneficiaries hospitalized under observation status have greater concerns about their cost-sharing obligations than their higher income peers. Cost-sharing for observation care may have unintended consequences on utilization for low-income beneficiaries. Future studies should examine this potential relationship on a larger scale.

Peer Review reports


Observation status or “hospital outpatient status” is a classification for Medicare beneficiaries that are billed as outpatients for a hospitalization. Whereas hospital inpatients are billed through Medicare Part A, observation patients are billed through Medicare Part B, which can result in higher out-of-pocket costs [1]. Since 2013, The Centers for Medicare and Medicaid Services (CMS) has defined observation patients by the 2-Midnight Rule, which stipulates that patients with an anticipated length of stay of < 2 midnights be designated as observation status while those anticipated to require ≥2 midnights be designated as inpatients, regardless of clinical status. Prior to this, observation status was determined by clinical criteria.

Observation visits have been on the rise for over a decade. Between 2006 and 2010, the use of observation stays increased 70%, largely as a result of penalties that CMS was imposing on hospital systems under the Recovery Audit Contractor Program for inappropriately billing for short-stay admissions [2, 3]. The introduction of the 2-Midnight Rule has increased the use of observation stays by another 8% [4]. It is estimated that approximately 25% of adult general medicine hospitalizations are observation visits [5].

Observation patients and inpatients have different cost-sharing responsibilities. Under Medicare Part A, inpatients pay a substantial deductible ($1364 in 2019), which covers most of the care in the hospital, 21 days in a skilled nursing facility if needed after the hospital stay, and hospital readmissions within 60 days of discharge [6]. Observation patients have a lower deductible, ($185), and higher cost-sharing for hospital services (20%) under Medicare Part B. They also are billed for “self-administered medications” which are medications that are not directly related to the primary diagnosis for the hospitalization [7]. Unlike Part A, Part B does not cover the cost of rehospitalizations, so patients can face cumulative out-of-pocket expenses for repeat observation stays. A prior study of Medicare claims demonstrated that the median out-of-pocket costs per observation stay are $448.94 [8] and that cumulative out-of-pocket expense for re-hospitalizations for observation stays exceeds that of the Medicare Part A deductible for over 25% of beneficiaries [9].

Our prior work has demonstrated that low-income patients are at higher risk for higher utilization of observation care and higher out-of-pocket costs related to this care compared to higher income patients [10]. Additionally, there is an association between cost-sharing and healthcare rationing among low-income Medicare beneficiaries [11, 12]. As more beneficiaries are exposed to Part B cost- sharing for their hospitalizations it is unclear whether this will impact beneficiary attitudes and behavior surrounding observation care. Our objective was to explore whether cost-sharing for observation care could impact future medical decision-making related to such care among low-income versus higher-income Medicare beneficiaries.


This is a pilot study of a convenience sample of Medicare beneficiaries hospitalized under observation status to Christiana Care Health System (CCHS) from January 4, 2016-May 19, 2016.


Eligible patients had traditional Medicare as their primary form of insurance and were hospitalized under the care of one of the non-teaching hospitalist services at either Wilmington Hospital, a 241 bed community hospital in Wilmington, DE or Christiana Hospital, a 913 bed tertiary care facility in Newark, DE. Patients were excluded if they were younger than 18 years of age, ICU or step-down admissions, direct admissions from home, transfers from other services, on the teaching service, listed as “comfort care only”, or as a “confidential” patient. Designated family members could answer questions at the patients request or if the patient was confused or had a history of significant cognitive impairment. The CCHS health informatics team sent a daily list of eligible patients to the principal investigator and research nurse. All eligible patients were approached during weekday, workday hours. The convenience sample was obtained based on patient availability, willingness, and ability to complete the survey.

Survey instrument

A 23-item survey was administered in person by a trained, unblinded, research assistant, within 24 h of their hospitalization. The first 3 questions addressed awareness and understanding of observation status and policies and were based on language from a CMS patient information pamphlet [7].

Questions 4–9 and 11–16, were taken directly from the 2014 National Health Interview Survey (NHIS) and addressed rationing of health services (mental health, vision care, dental, and specialist and primary care) and prescriptions due to cost in the previous 12 months. These questions were chosen to examine baseline concerns related to health care cost-sharing. The original NHIS questions had 4 responses (Yes, No, Refused, Don’t Know); We modified this slightly by offering the responses: Yes, No, Refused, N/A. Any patient who had a positive (yes) response to any of the NHIS questions related to rationing was coded has having health-related cost concerns. Question 10, was a modified version of NHIS question AAU.113 regarding concern about paying for the current hospitalization [13], and was analyzed as a dichotomous variable. Questions 17–21 were demographic in nature (age, race, ethnicity, self-reported income,). Income above or below 200% of the federal poverty line was assessed (approximately 31 K for a family of 2) to define low-income patients [14] (Additional file 1).

The last 2 questions assessed whether comprehension of Medicare observation policy could impact future health decision-making. Question 22 asked whether the patient would inquire about observation status on a future admission and had a yes/no response. Question 23 asked whether the patient would: 1) stay for the care prescribed in the hospital 2) ask the provider to arrange for services as an outpatient or 3) leave against medical advice (AMA) if hospitalized under observation status again (Additional file 1). Prior to asking questions 22 and 23, the research assistant reviewed the answers to questions 2&3 that addressed specifics of the Medicare observation policy to ensure that the patient understood the cost-sharing implications. Data regarding date of birth, age, insurance and hospitalist group were obtained from the electronic medical record. Race and ethnicity data were obtained from the electronic medical record for non-respondents. All responses were recorded on a paper copy and entered into REDCap software.


Descriptive statistics were reported for demographics, baseline health care rationing behaviors, and understanding of observation policies. Chi square was used to determine sociodemographic differences between responders and non-responders and to determine whether income level was associated with a history of health care rationing and cost concerns and future decision-making related to observation care. All statistical analysis was performed using STATA13, College Station, TX. All patients and/or the designated family member signed informed consent. The Christiana Care Health System Institutional Review Board approved this study.

Sample size

Based on prior work that demonstrated a 3–4% reduction in health care utilization after introduction of cost-sharing measures [11, 12], we estimated that 686 interviews were needed to achieve 80% power for a two-tailed test with 95% confidence. For this pilot study, we had resources to perform interviews for 6 months.


There were 466 patients eligible for the study, and 168 (38%) were either missed or discharged prior to being approached for enrollment. Of those approached, 127 had a medical or cognitive impairment that prohibited completion and 11 refused. There were 160 patients who completed interviews, and of these, 150 were completed by the patient, and 10 by designated family members. Post-hoc analysis determined that 16 respondents did not have traditional Medicare as their primary insurance and were excluded, leaving 144 for analysis (Additional file 2: Figure S1). Respondents were more likely to be younger than non-respondents (p = 0.008) but were otherwise similar (Table 1).

Table 1 Characteristics of Survey Respondents and Non-Respondents

Close to half (45.8%; 66/144) of respondents were low-income by self-report. Among those who responded, approximately 17% (16.8%;24/143) did not have supplemental insurance, 37.1% (36/97) delayed seeking care due to cost concerns and 42.5% (51/120) rationed prescriptions due to cost in the previous 12 months.

Over half of respondents (52.8%; 76/144) were aware of their observation status but only 8.8% (11/125) answered both questions related to observation policies correctly (Table 1). Approximately 1/3 (34.4%; 41/119) of respondents expressed concern about paying for their observation stay. Low-income respondents were more likely to be concerned about the cost of their stay than higher-income respondents (70.7% vs 29.3%, p = 0.015) (Table 2).

Table 2 Association between Income level, Cost concerns, and Future Utilization

Potential impact of observation cost-sharing on future observation utilization

The majority of respondents (72.3%; 68/102) stated that they would stay for the duration of the hospitalization if they were hospitalized under observation status in the future. However, approximately 1/3 (34%; 32/102) would request that their work-up be performed as an outpatient. Low-income beneficiaries were less likely to stay for the care they needed (48.5% vs 51.5%), more likely to request outpatient completion of their workup (56.3% vs 43.8%), and more likely to leave AMA if hospitalized under observation status again (100% vs 0%), though these trends were not statistically significant (Table 2).


In this pilot study of 144 Medicare beneficiaries hospitalized under observation status, most patients were aware of their status, but very few understood the cost-sharing implications. Low-income beneficiaries were more concerned about the cost of their hospital stay and were more likely to consider requesting outpatient completion of their work-up or leaving AMA compared to higher-income beneficiaries. Our findings imply opportunities for early or pre-admission education of Medicare beneficiaries regarding observation status, and consideration of the potential unintended consequences of the current policies.

From a legal standpoint, Medicare beneficiaries are required to be made aware of their cost-sharing responsibilities under observation status [15]. However it is still unclear how the cost-sharing itself may impact future behavior. In the outpatient setting, prior work has demonstrated that low-income patients are more likely to defer, delay or ration care if their cost-sharing responsibilities increase [11, 12, 16]. This may be associated with higher hospital utilization [12].

Although the cost-sharing responsibilities related to observation care have been in place for decades, the number of patients exposed has increased [3, 17]; as of 2010, 9.3% of beneficiaries were hospitalized under observation status [1] and the number is expected to rise [18]. Medicaid can provide significant protections against out-of-pocket costs as can supplemental insurance programs, such as Medigap; However, many patients who are eligible for Medicaid do not enroll [19] and Medigap programs, can be costly [20, 21]. Nationally, as was reflected in our study sample, approximately 14% of beneficiaries, representing over 6,700,000 individuals are not enrolled in supplemental insurance plans [21]. These patients often do not qualify for Medicaid but cannot afford supplemental coverage and are at greatest risk for high out-of-pocket costs [21].

In the context of our limited study, we found that low-income beneficiaries were more likely to express concern about the cost of their hospital stay and were more likely to consider outpatient work up or leaving AMA if hospitalized under observation status in the future compared to higher-income beneficiaries. We also found that many beneficiaries already ration health care services, a sign of significant financial strain [11, 12, 16]. As observation hospitalizations continue to rise, it will be important to proactively identify and support beneficiaries at risk for significant health care cost burden. Future research should prospectively evaluate health utilization behaviors of patients with significant out-of-pocket costs related to observation care.


There are several important limitations to our study. We surveyed a convenience sample of beneficiaries, which limits the generalizability of our findings. Additionally, a substantial number of eligible patients were excluded because they were either confused/demented for discharged within 24 h. This could have biased our sample in favor of sicker observation patients, without cognitive impairment, which also limits the generalizability of our findings. In the context of this pilot study we were underpowered, and were only able to detect a potential signal regarding the direction of our findings. However, our initial sample size calculation was based on an estimate that 3–4% of patients would ration observation care due to concerns related to cost-sharing, and our study found that a substantially higher proportion would consider forgoing care related to their observation stay after learning about cost-sharing obligations (34% would request outpatient work-up and 2% would leave AMA). Assuming our sample was generalizable, a larger patient sample could further strengthen the association that we found. However, a large study using a more representative patient sample could potentially result in a smaller effect size. An additional limitation was that the questions that addressed patient understanding of observation policy and the potential impact on future decisions were not validated. We could not identify any validated surveys that addressed these questions. However, we tested our survey on 10 Medicare observation patients for comprehension and not verified.


As exposure to cost-sharing for observation care increases, its impact on the health behaviors, financial, and physical wellness of our most vulnerable beneficiaries remains unclear. Future prospective studies should examine this issue on a larger scale.



Against Medical Advice


Christiana Care Health System


Centers for Medicare and Medicaid Services


National Health Interview Survey


  1. Kangovi S, Cafardi SG, Smith RA, Kulkarni R, Grande D. Patient financial responsibility for observation care. J Hosp Med. 2015;10(11):718–23.

    Article  PubMed  Google Scholar 

  2. Wachter RM. Observation status for hospitalized patients a maddening policy begging for revision. JAMA Intern Med. 2013;173(21):1–2.

    Article  Google Scholar 

  3. Baugh CW, Schuur JD. Observation care - high value care or a cost-shifting loophole? Nejm. 2013;369(4):302–5.

    Article  CAS  PubMed  Google Scholar 

  4. Wright B, O’Shea AMJ, Ayyagari P, Ugwi PG, Kaboli P, Sarrazin MV. Observation rates at veterans’ hospitals more than doubled during 2005-13, similar to medicare trends. Health Aff. 2015;34(10):1730–7.

    Article  Google Scholar 

  5. Sheehy AM, Graf B, Gangireddy S, et al. Hospitalized but not admitted. JAMA Intern Med. 2013;53705(21):1991–8.

    Article  Google Scholar 

  6. Centers for Medicare and Medicaid Services. Your Medicare Coverage Medicaregov Available at: Accessed January 1, 2016.

  7. Centers for Medicare and Medicaid Services. Are you a hospital inpatient or outpatient? C Prod No 11435. 2014. Available at: Accessed 1 Jan 2016.

  8. Goldstein JN, Schwartz JS, McGraw P, Hicks LS. Implications of cost-sharing for observation care among Medicare beneficiaries. Society of General Internal Medicine, Annual Meeting, Poster Presentation April 2018.

  9. Kangovi S, Cafardi SG, Smith RA, Kulkarni R. Patient Financial Responsibility for Observation Care. 2015;10(11). doi:

  10. Goldstein JN, Zhang Z, Schwartz JS, Hicks LS. Observation status, poverty, and high financial liability among Medicare beneficiaries. Am J Med. 2017.

  11. Trivedi AN, Rakowski W, Ayanian JZ. Effect of cost sharing on screening mammography in Medicare health plans. N Engl J Med. 2008;358(4):375–83.

    Article  CAS  Google Scholar 

  12. Trivedi AN, Moloo H, Mor V. Increased ambulatory care copayments and hospitalizations among the elderly. N Engl J Med. 2010;362(4):320–238.

    Article  CAS  Google Scholar 

  13. National Center for Health Statistics. National Health Interview Survey. 2014. Available at: Accessed January 1, 2016.

  14. Mccormack LA, Gar SA, Hibbard JH, Keller SD, Kilpatrick KE, Kosiak B. Health insurance knowledge among Medicare beneficiaries. Health Serv Res. 2002;37(1):41–61.

    Article  Google Scholar 

  15. US Congress. Public Law 114–42 114th Congress An Act: NOTICE ACT. US Congress; 2015:468–469. Available at:

  16. Goldstein JN, Long JA, Arevalo D, Ibrahim SA. US veterans use vitamins and supplements as substitutes for prescription medication. Med Care. 2014;52(12):65–9.

    Article  Google Scholar 

  17. Levinson, Daniel R. Vulnerabilities remain under Medicare’s 2-midnight policy.; 2016. Available at:

  18. Sheehy AM, Caponi B, Gangireddy S, et al. Observation and inpatient status: clinical impact of the 2-midnight rule. J Hosp Med. 2014;9(4):203–9.

    Article  PubMed  Google Scholar 

  19. Gross D, Alecxih L, Gibson M, Corea J, Caplan C, Brangan N. Out-of-pocket health spending by poor and near-poor elderly Medicare beneficiaries. Health Serv Res. 1999;34(1):241–54.

    CAS  PubMed  PubMed Central  Google Scholar 

  20. Centers for Medicare and Medicaid Services. Costs of Medigap Policies. Available at:

  21. Jacobson G, Neuman T, Damico a. Medigap Enrollment Among New Medicare Beneficiaries.; 2015. Available at:

Download references


We would like to thank Kaitlyn Sternat, BA for her assistance in data collection.

Availability of data and materials

Data/materials is available from Dr. Goldstein on request.

Ethics approval and consent to participate*

The Christiana Care Institutional Review Board approved this study. All study subjects provided verbal informed consent as the research presented no more than minimal risk of harm to subjects, and this was approved by the Christiana Care Institutional Review Board. Designated family members could answer questions at the patients request or if the patient was confused or had a history of significant cognitive impairment. This was approved by the Christiana Care Institutional Review Board.

Author information

Authors and Affiliations



Study Design: JNG, PM, JSS, LSH. Data Collection: JNG, PM. Analysis: JNG, PM, JSS, LSH. Manuscript Drafting/editing: JNG, JSS, LSH. All authors have read and approved the manuscript.

Corresponding author

Correspondence to Jennifer N. Goldstein.

Ethics declarations

Consent for publication

Not applicable.

Competing interests

Dr. Schwartz has received consultancy fees from Bayer, Pfizer, Takeda, The Blue Cross Blue Shield Association, and the Agency for Healthcare Research and Quality, unrelated to the content of this study. All other authors declare no conflicts of interest.Funding: Dr. Hicks was supported by an Institutional Development Award (IDeA) from the National Institute of General Medical Sciences of the National Institutes of Health under grant number U54-GM104941 (PI Binder-Macleod). The funding source had no role in the design of the study, data collection, analysis, interpretation of data or writing the manuscript.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Additional files

Additional file 1:

Appendix 1. Patient Interview. (PDF 48 kb)

Additional file 2:

Figure S1. Study Sample. Characteristics of respondents and non-respondents. (TIF 58 kb)

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Goldstein, J.N., Schwartz, J.S., McGraw, P. et al. “Implications of cost-sharing for observation care among Medicare beneficiaries: a pilot survey”. BMC Health Serv Res 19, 149 (2019).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: