The Assessing Care of Vulnerable Elders (ACOVE) project has developed several quality indicator sets for ambulatory geriatric and nursing home patients (5, 6). These quality frameworks, however, cannot be simply adopted. Home-based primary care patients have higher mortality rates and shorter life expectancies than the ACOVE-studied populations, with goals of care that focus on quality, rather than prolongation, of life (7–9). Home-based primary care programs are often multidisciplinary efforts that emphasize coordination and continuity of care (10). Furthermore, processes of care vital to high-quality care of the homebound patient may have been overlooked by the ACOVE researchers. Given these findings, the Home-based Primary Care Quality Initiative (HPCQI), a multistep, national expert panel process, was completed. The study, by adapting and expanding on earlier ACOVE work, identified a set of evidence-based process quality indicators that are valid and important to providing high-quality primary care to homebound seniors.

Methods

Top Methods Results Discussion Author & Article Info References

Setting

The study was conducted at the Mount Sinai Visiting Doctors program of the Mount Sinai Medical Center, New York, New York (11). The initiative was formally endorsed by the American Academy of Home Care Physicians (AAHCP) Board of Directors in 2005.

ACOVE Overview

The ACOVE project created a comprehensive quality-of-care management program for frail community-dwelling elders (5), including an evidence-based quality indicator set covering 22 geriatric conditions (12). The feasibility of using these indicators has been demonstrated (13).

HPCQI Overview

The HPCQI team used a multistep process for adapting and expanding the ACOVE quality indicators to home-based primary care. First, the study team modified language in the ACOVE community and nursing home quality indicators for the home care setting and eliminated hospital-based quality indicators. Second, the HPCQI team members developed evidence-based quality indicators for constipation and insomnia, 2 conditions that the ACOVE project did not cover but for which homebound patients have disproportionate morbidity. Third, a home-based primary care quality indicator development expert panel (development panel) assessed the new and modified quality indicators for applicability to the home-based primary care setting. Fourth, a second, discrete panel—the home-based primary care quality indicator evaluation expert panel (evaluation panel)—evaluated the applicable quality indicators for validity and importance. Finally, indicators that were deemed to be valid and important by the evaluation panel were accepted for use in the home-based primary care setting.

To form the panels, we contacted clinicians who were in home-based primary care program administration and quality management, had authored peer-reviewed scholarship in the area of home-based primary care, or had leadership roles within the AAHCP. A group of panelists that were diverse in practice type, location, and setting was a priority. (The Appendix lists the members of each panel.)

The development panel's primary task was to perform a face validity review through a mailed survey of the ACOVE quality indicators. Panelists rated the indicators as 1) likely to apply equally to the home-based primary care setting as to the nursing home or ambulatory care setting, 2) likely to need modification, or 3) unlikely to apply.

The evaluation panel rated the validity and importance of the applicable quality indicators through 2 rounds of mailed surveys. Panel members used a modified Delphi expert panel group judgment process, which has been shown to be appropriate for translating research from 1 area or study to a different population or setting (14, 15). Our study substituted conference calls and mailed surveys for the face-to-face approach—a modification that is shown to have acceptable reproducibility (16).

Throughout the study, panel members used widely accepted definitions of "home-based primary care" (17) and "homebound seniors" (18).

Role of the Funding Sources

This project was supported by funds from the Division of State, Community, and Public Health, Bureau of Health Professions, Health Resources and Services Administration, U.S. Department of Health and Human Services and a Geriatric Academic Career Award (Dr. Boal). The funding sources played no role in the design, process, or interpretation of the study or in the decision to submit the manuscript for publication.

Results

Top Methods Results Discussion Author & Article Info References

Quality Indicator Development Panel

Of the 260 quality indicators evaluated by the development panel, 25 (10%) were rated as inapplicable and were removed from the quality indicator set.

Quality Indicator Evaluation Panel

On the basis of first-round survey results, 19 (8.1%) of the 235 applicable indicators, rated as both invalid and unimportant, were removed from the quality indicator set. One hundred seventy-six indicators, rated as valid and important, were included in the final quality indicator set. Forty (17.0%) indicators received indeterminate evaluations, were discussed by the panelists during 2 conference calls, and were then rerated. Of these rerated indicators, 24 were included and 16 were removed from the final quality indicator set.

Finally, the evaluation panel reviewed the indicators that the development panel eliminated and agreed that they were appropriately excluded.

Final Home-Based Primary Care Quality Indicator Set

The result is a quality indicator set with 200 process indicators for 23 geriatric conditions (Appendix Table). Of these 200 quality indicators, 21 (10.5%) were newly created, 52 (26%) were modified from the ACOVE work, and 127 (63.5%) were unchanged from the ACOVE work. Fifty-five ACOVE indicators and 5 newly created indicators, rated as inapplicable, invalid, and/or unimportant, were dropped from the quality indicator set.

The 21 new indicators were limited to 6 conditions: constipation (n = 9), continuity and coordination of care (n = 3), end-of-life care (n = 2), heart failure (n = 1), insomnia (n = 5), and preventive care (n = 1). The 52 modified indicators fell into 6 categories: documentation (n = 5); earlier screening, intervention, and follow-up (n = 19); more specific or explicit (n = 9); more specific to the home-based primary care setting (n = 6); promoting palliative care or patient autonomy (n = 6); and miscellaneous (n = 7) (Table 1).

Tables 2 and 3 show selected characteristics of the evaluated indicators. According to domain of care, the highest percentage of rejected indicators was treatment-related (51.7%). Accepted indicators were more evenly dispersed among the 4 domains of care.

Discussion

Top Methods Results Discussion Author & Article Info References

Home-based primary care experts identified a core set of 200 processes that all home-based primary care practitioners should provide for their patients. For our study, we pursued adaptation and expansion of the ACOVE work, rather than adoption or de novo construction. The result is a quality indicator set that overlaps the well-established ACOVE framework with modifications for applicability to the home-based primary care setting.

The modifications, deletions, and additions made during the study demonstrate the unique challenges and preferences of homebound seniors (7–10). More than half of the eliminated indicators were from the treatment domain of care. Thus, treatment quality indicators make up a smaller percentage of the overall quality indicator set than in the original ACOVE work. Quality indicators for follow-up and continuity of care domains, on the other hand, have a higher proportional representation in the quality indicator set. The quality indicator modifications trended toward closer management. Reflecting the reality that homebound patients can decline quickly, many modifications were for more frequent screening and follow-up. Finally, the bulk of the new indicators were in quality-of-life domains, such as end-of-life care, constipation, and insomnia, further reinforcing the primacy placed on patient comfort.

Our most vulnerable seniors are increasingly relying on home-based primary care. Infrastructure to support the provision of high-quality home care is lacking. As such, our quality indicator set fills an essential need. These indicators should guide home-based primary care practitioners as they look to provide high-quality care.

In addition to improving quality, our quality indicator set will allow for future comparative research. An important question, yet unanswered, about home-based primary care is whether the model provides care that is similar in quality to that of more established systems. Without quality-of-care data, the medical community and government payers will probably continue to resist embracing home-based primary care. By creating a quality instrument that has substantial overlap with work done in other geriatric care settings, comparative quality of care can be studied.

Feasibility of implementation could be a limitation to our study. Home-based primary care programs are typically resource-poor and might have difficulty funding quality management infrastructure. Despite these feasibility concerns, all valid and important indicators were reported to encourage home-based primary care programs to search for innovative care solutions and novel funding sources. Future projects to decrease the cost of implementing the quality indicator set should be pursued.

Many home-based primary care patients have severe dementia or fewer than 6 months to live. These patients have different goals of care that render many processes inappropriate. Future studies, such as that completed by ACOVE researchers (19), should identify quality indicators that are not appropriate for these patients.

Another concern, common to expert opinion processes, is that many of these indicators need validation through prospective study. A process for regular update and review of this quality instrument should be coupled with prospective validation. Finally, our quality indicator set represents the expert opinion of a few home-based primary care practitioners. The diversity of the panelists and that they are recognized national experts suggests, however, that their opinions are broadly representative.

Homebound seniors typically have several chronic medical conditions that make providing high-quality primary care both critical and difficult. Using a multistep process, home-based primary care experts identified processes of care that are essential to providing high-quality care to homebound seniors. This quality indicator set should provide a rigorous structure for home-based primary care programs to build their quality-of-care management programs and should improve patient quality of life and longevity. The overlap of the set with previous ACOVE work also allows for future comparative study of quality provided by the home-based primary care model.

Appendix: Panel Members

Development Panel

Lisa Caruso, MD, MPH, Boston University School of Medicine, Boston, Massachusetts; Eric DeJonge, MD, George Washington University School of Medicine, Washington, DC; Kevin Jackson, MD Geriatric Solutions, Phoenix, Arizona; Deirdre Mole, RN, GNP, Weill Medical College of Cornell University, New York, New York; Jean Yudin, RN, CMSN, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania.

Evaluation Panel

Peter Boling, MD, Virginia Commonwealth University, Richmond, Virginia; Thomas Cornwell, MD, Central DuPage Hospital, Wheaton, Illinois; Jennifer Hayashi, MD, Johns Hopkins Bayview, Baltimore, Maryland; Benneth Husted, DO, Housecall Providers Inc., Portland, Oregon; Sharon Levine, MD, Boston University School of Medicine, Boston, Massachusetts; Veronica LoFaso, MD, Weill Medical College of Cornell University, New York, New York; Sonni Mun, MD, Mount Sinai School of Medicine, New York, New York; Wayne McCormick, MD, University of Washington, Seattle, Washington; Edward Ratner, MD, University of Minnesota, Minneapolis, Minnesota.