Sarah M. Donelson, M.A., Christopher M. Murtaugh, Ph.D., Penny Hollander Feldman, Ph.D., Kamal Hijjazi, Ph.D., Lori Bruno, M.P.H., Stephanna Zeppie, R.N., M.A., C.P.H.Q., Shiela Kinatukara Neder, R.N., B.S.N., Eva Quint, B.A., R.N., B.S.N., Liping Huang, M.A., and Amy Clark, B.A.
Center for Home Care Policy and Research
Visiting Nurse Service of New York
This report was prepared under contract #100-99-0020 between the U.S. Department of Health and Human Services (HHS), Office of Disability, Aging and Long-Term Care Policy (DALTCP) and the Visiting Nurse Service of New York. For additional information about this subject, you can visit the DALTCP home page at http://aspe.hhs.gov/_/office_specific/daltcp.cfm or contact the Office at HHS/ASPE/DALTCP, Room 424E, H.H. Humphrey Building, 200 Independence Avenue, S.W., Washington, D.C. 20201. The e-mail address is: webmaster.DALTCP@hhs.gov. The Project Officer was Kamal Hijjazi.
The Principal Investigator for this project was Sarah M. Donelson, M.A., from September 1999 through May 2000. She is now with PeaceHealth in Eugene, Oregon. The Principal Investigator since June 2000 has been Christopher M. Murtaugh, Ph.D.
The opinions and views expressed in this report are those of the authors. They do not necessarily reflect the views of the Department of Health and Human Services, the contractor or any other funding organization.
The Social Security Act -- Sections 1814(a)(2)(C) and 1835(a)(2)(A), as amended by Section 4615 of the Balanced Budget Act of 1997 -- establishes the basic eligibility and coverage requirements for Medicare home health benefits. The act states that home health services shall be provided to beneficiaries who (1) are homebound, (2) have medical necessity (i.e., need intermittent, skilled nursing care, physical therapy, speech therapy or continued occupational therapy), and (3) are under a physician’s plan of care. While these criteria may seem straightforward, their application is somewhat difficult. In fact, both the use and interpretation of the first two criteria vary widely among Medicare Fiscal Intermediaries (FIs), home health agencies and physicians (Office of the Inspector General (OIG) Report A-04-94-02087).
How the homebound and medical necessity criteria can be interpreted in different ways becomes apparent when the language in the Health Care Financing Administration’s (HCFA) home health coverage guide (HIM-11) is examined. The HIM-11 states that a person may be considered homebound if leaving the home requires considerable and taxing effort. Absences from the home are acceptable, provided they are infrequent, of short duration or to receive medical treatment. Words such as “taxing”, “considerable”, “infrequent” and “short” may have very different meanings depending on the interpreter since there is no objective and measurable definition of these terms. The language used to describe medical necessity is similar. Section 205 of the HIM-11 defines medical necessity using words and phrases such as “complexity,” “safely and effectively,” “trained,” “able and willing” and “unique medical condition.”
Congress, in Section 4613(a) of the Balanced Budget Act of 1997, charged the Secretary of the Department of Health and Human Services (HHS) to investigate the criteria for determining homebound status and to develop a method of applying such criteria (Shalala, 1999). The study was completed by HHS and a report was submitted to Congress in 1999. The findings from the study were inconclusive. However, the report recommended that there be further examination of the issue once more comprehensive data sources were available (e.g., the Outcome Assessment and Information Set (OASIS)). The above concerns and recommendations along with recommendations from the OIG (1999) and the Medicare Payment Advisory Commission (1999) prompted the Office of the Assistant Secretary for Planning and Evaluation (ASPE) within HHS to study how to clarify the homebound and medical necessity criteria through the use of OASIS data.
This study assesses the feasibility of using information routinely collected as part of OASIS, as well as other patient data, to develop objective and consistent tools for evaluating a beneficiary’s homebound status and his or her need for skilled care under the Medicare home health benefit. In the first phase of the project we developed and tested two OASIS-based algorithms. Subsequently, we developed and applied medical record review tools for assessing whether patients meet the homebound and medical necessity criteria.
In December 2000, just as this project was concluding, new legislation affecting the homebound definition was passed by Congress and signed into law. Section 507 of the Beneficiary Improvement and Protection Act amends the Social Security Act and clarifies the homebound definition under the Medicare home health benefit by expanding the list of circumstances in which absences from the home are consistent with a determination that a beneficiary is homebound. Specifically, the law allows absences from the home attributable to the need to receive health care treatment from an adult day-care program as well as absences for religious services. While this recent change in the law adds to the numerous examples in the HIM-11 concerning when beneficiaries meet the homebound and medical necessity criteria, the examples are not exhaustive enough to greatly reduce the need for interpretation when FIs, providers and beneficiaries assess eligibility.
There still is a need, then, for objective and consistent tools that increase overall understanding of the home health benefit, and reduce problems and frustration experienced at the agency and beneficiary level due to wide variation in interpretation of the rules. FIs, for example, would have a single standardized method for applying the homebound and medical necessity criteria and would be able to focus their reviews on more complex cases, saving both time and effort. Physicians would be able to use the tools to improve their understanding of the types of patients who qualify for the home care benefit. Home care providers could use OASIS to determine which patients are highly likely to meet the two Medicare eligibility requirements and when documentation concerning a patient’s homebound status or need for skilled services is particularly important. While such tools will not eliminate the need for judgment, particularly in difficult or “gray” areas, they have many potential uses and benefits.
We conducted a comprehensive review of government, association and academic literature on homebound individuals and medical need for skilled care in the first weeks of the project. Subsequently, national experts in home care were identified and a panel of nine clinicians convened in New York City. The expert panel was asked to identify clinical, functional and other reasons why individuals are homebound and in need of skilled nursing or therapy services. They also provided advice on whether these reasons can be proxied using OASIS data. The project team analyzed the results of the literature review and expert panel and developed two OASIS-based algorithms and two medical record review tools.
The two algorithms and medical record review tools were tested using data from six home care agencies with considerable OASIS experience. The agencies selected to participate in the study varied in geographic location, Medicare FI, size and ownership. Each agency submitted OASIS data from the last quarter of 1999 (October-December). A sample of 100 start-of-care assessments was selected from each agency’s pool of Medicare patient assessments for beneficiaries who were not enrolled in a Medicare health maintenance organization, for a total sample size of 600 Medicare beneficiaries. The algorithms then were applied to determine the number of patients classified as homebound and in need of skilled nursing or therapy services based solely on OASIS data.
The final phase of the project was a medical record review of the records of patients who were not captured by either of the two algorithms. Two medical record review tools were constructed (one for the homebound and one for the medical necessity criterion). Each tool included a check-off list of clinical conditions identified by the expert panel as important for determining eligibility but OASIS data are lacking. Nurses with extensive chart review experience went through each patient record, checked off items on the tool when documented in the record, and made a professional judgment about whether the beneficiary met the homebound and/or medical necessity requirements.
The results of this project indicate that OASIS information can be used to identify patients highly likely to meet the homebound and medical necessity criteria at the start of care (Table 1).
Overall, 48.5% of the 600 patients screened were classified as homebound based solely on information from OASIS included in the homebound algorithm. Roughly half of the patients met this eligibility criterion because they experienced shortness of breath with minimal or moderate exertion. One-third were classified as homebound because of mobility limitations.
A much higher proportion of the 600 patients screened (89.0%) met the medical necessity criterion based solely on information from OASIS included in the second algorithm. Almost half of the patients meeting the criterion did so because primary diagnosis symptoms were poorly controlled, or controlled with difficulty, in combination with other reasons for skilled care. One-quarter met the criterion because they had a pressure ulcer, stasis ulcer or surgical wound.
Medical record review then was conducted on 322 records of patients who were not captured by one or both of the algorithms. The large majority of the reviews were for patients who were not captured by the homebound algorithm. Over half of these patients were judged by a clinician to meet the Medicare homebound criterion based on a detailed chart review. Similarly, a clinician concluded that half of the much smaller number of patients who were not captured by the medical necessity algorithm met this Medicare requirement.
Two questions concerning completeness of OASIS data and patient records were included at the bottom of the medical record review tools (see Table 2). The data quality problems encountered by the nurses conducting the medical record reviews were substantial. Conflicting information was a common problem. The nurse reviewers were asked to indicate whether OASIS items were “obviously miscoded” and “yes” was checked in 45.9% of the homebound reviews and 52.4% of the medical necessity reviews. However, in many of these cases the nurse noted that there was a discrepancy between an OASIS item and another part of the patient record but they were unable to determine which of the two was correct. The nurses also reported relatively high levels of conflicting information in chart notes and gaps between notes (i.e., “inadequate charting”). (Missing OASIS items and missing pages or sections of patient records occurred relatively infrequently.) The data quality problems encountered by the nurses were relatively similar regardless of whether nurses were determining homebound status or the need for skilled care.
When the algorithm and medical record review results are combined, 77.0% of the 600 patients were classified as meeting the homebound criterion and 94.5% were classified as meeting the medical necessity criterion (Table 3). Agencies failed to send only 16 of the 322 records requested. The results would not change substantially, therefore, even if we had access to the missing records.
There are three important study limitations. First, the algorithms and medical record review tools only were tested on a small non-random sample of Medicare beneficiaries. How well the tools developed here will work on a more representative sample needs to be determined. Second, the study was limited to start-of-care assessments. It was beyond the scope of this project to test the algorithms and tools on patients remaining in home care beyond 60 days. Finally, the data analyzed were from the final quarter of 1999 -- the middle of a period of substantial change in the home care market. The extent to which the case mix of Medicare patients served by agencies has changed over the last 16 months is unclear.
There are two other potential limitations that need to be acknowledged. Nurses from a single agency conducted the medical record reviews. While they all had extensive chart review experience, the reliability of the medical record review findings is uncertain since a rigorous assessment of inter-rater reliability was beyond the scope of the project. In addition, the accuracy of the OASIS data used to test the algorithms was not formally assessed.
Conclusions and Recommendations
The results of this project show that the two OASIS algorithms successfully identify patients highly likely to meet the homebound and medical necessity criteria for Medicare home health care. Using OASIS data alone, almost 90% of the 600 patients in the sample were classified as meeting the medical necessity criterion. The figure for the homebound criterion is lower (48.5%). It is critical to note, however, that individuals may be clearly eligible for home care benefits even though they are not captured by the OASIS algorithms. Nurses who were experts in chart review, in fact, concluded that over half of the patients not captured by the homebound algorithm actually met this requirement based on a careful review of the patients’ medical records.
Overall, we believe that both the algorithms and medical record review tools have the potential to be very useful to Medicare FIs and providers (e.g., home care agencies and physicians) after additional testing and refinement. The algorithms can be used to ascertain whether an individual meets the two Medicare eligibility requirements and the record review tools to reinforce the importance of documenting the need for home care for patients who are not captured by the algorithms. Agencies, for example, could incorporate the algorithms into their routine OASIS data processing procedures and assess Medicare beneficiaries’ eligibility according to the algorithms and the adequacy of medical record documentation prior to OASIS submission to HCFA.
We recommend, as the next step in the development of the algorithms and medical record review tools, additional testing using a large, nationally representative sample of Medicare beneficiaries. We obtained data from a heterogeneous group of agencies with regard to geographic location, agency ownership, profit status, rural/urban location and size; nevertheless, the six agencies in this study are not representative of all home care agencies. It is important to test the tools developed in this project on a representative sample to ensure that the results are the same for the country as a whole. In addition, we recommend that the algorithms and medical record review tools be tested at the point that patients are recertified for Medicare home health care (i.e., at 60-day intervals following admission to home care). While we believe the algorithms and medical record review tools will work well with follow-up assessments, this remains to be empirically tested.
We also recommend further research to improve the “sensitivity” of the homebound algorithm (i.e., the ability of the algorithm to capture individuals who truly are homebound) and assess the specificity of both algorithms (i.e., the ability of the algorithms to exclude individuals who do not meet the criteria). A large proportion of the patients judged to be homebound by nurses based on chart review had “general weakness or somnolence.” Our results suggest that with further work it may be possible to identify at least some of the individuals with these conditions who are homebound using a combination of OASIS items that could be added to the algorithm. In general, more research is needed to refine the homebound algorithm to increase the percent of individuals correctly classified. Research also should be conducted to determine the extent to which the OASIS algorithms capture “false positives.” The recommendations of an expert panel of home care clinicians were followed when we constructed the OASIS algorithms. It still is possible, however, that some patients meeting the eligibility criteria according to the algorithms could be judged not to be homebound or in need of skilled care when their medical records are carefully reviewed. The extent of any such false positives should be determined as well as the reasons for false positives. Depending on the results, the algorithms should be refined to address this problem.
HIM-11, Chapter II -- Coverage of Services. Health Care Financing Administration.
Medicare Payment Advisory Commission. Access to Home Health Services. In, Report to the Congress: Selected Medicare Issues, June 1999. (pages 105-16)
OIG Report. Review of Medicare Home Health Services in California, Illinois, New York, and Texas (A-04-99-01194) October 1999.
OIG Report. Results of the Audit of Medicare Home Health Services in Florida (A-04- 94-02087) June 1995.
Shalala DE. “Homebound” -- A Criterion for Eligibility for Medicare Home Health Care. Report to Congress, April 1999.
|TABLE ES-1: Percent of Patients Meeting Each Criterion Using Only OASIS Data|
|TABLE ES-2: Response to Data Quality Questions|
|“Yes” Response to Question #1: Were OASIS start-of-care assessment items obviously miscoded or missing?1||136||45.9%||33||52.4%|
|“Yes” Response to Question #2: Was the patient record clearly incomplete?2||120||40.5%||34||54.0%|
|Both Questions Checked “Yes”||79||26.7%||25||39.7%|
|Either Question Checked “Yes”||177||59.8%||42||66.7%|
|NOTE: A total of 296 charts were reviewed to determine homebound status and 63 to determine whether the medical necessity criterion was met. |
|TABLE ES-3: Summary of Results|
|OASIS Algorithm Results|
|Criterion not met||309||51.5%||66||11.0%|
|Record Review Results|
|Criterion not met||125||20.8%||30||5.0%|
|No record received||13||2.2%||3||0.5%|
|Criterion not met||125||20.8%||30||5.0%|
|No record received||13||2.2%||3||0.5%|