6.1 What is the estimated magnitude of savings from more appropriate medical care for frail elders in a MediCaring Community?

The costs of overuse of medical interventions show up in virtually every effort to optimize care. An overview of the main innovations reported in the literature is in Chapter 3, Table 3.1: Evidence-Based Clinical Improvement Examples, on pages 57-66. As described in Chapter 3, a number of elder care reform programs, such as Independence at Home, PACE, GRACE, and the Veterans’ Home-Based Primary Care, have improved quality of life while substantially reducing use of hospitals and nursing homes and lowering costs.

The public often assumes that the waste in medical care is in high-tech interventions, like resuscitation, intensive care units, and transplants. In Medicare, the waste and very low value care is at least as likely to be in thoroughly ordinary care for routine situations, in which the elderly person is given ill-advised diagnostic testing or prevention screening as well as unwanted and often unused treatments. For example, my mother compressed a vertebra when she was 90 years old, probably when picking up a small suitcase. She ended up with two trips to the emergency room to get a diagnosis, multiple imaging studies (bone scan, CT scan, MRI scan, two sets of plain x-rays), a specially manufactured corset which was entirely too large, and an orthopedist who very much wanted to inject her spine to “reinflate” the vertebra (a procedure known as vertebroplasty). The cost of the tests and visits and corset came to about $10,000, and the vertebroplasty would have added at least $20,000 more. The vertebroplasty procedure has been the subject of two randomized controlled research trials (with the control group getting the identical patient experience without the actual injection, so no patients knew whether they got the actual intervention). Both studies showed that the vertebroplasty had no advantage, had some serious but uncommon immediate side effects, and was actually likely to increase the risks of more fractures of neighboring vertebrae.[160],[161],[162]

What did she need for optimal care? No diagnostic test or medical treatment has been shown to improve upon one ordinary X-ray to confirm the diagnosis and evaluate the spine generally, followed by pain relief and gradual return to function. This requires support by a skilled home care nurse to manage pain medications, encourage re-ambulation, assure nutrition, and arrange homemaker services as needed over the next month or two. All those imaging studies were searching for a cancer, but they could have been done for the tiny percent of elderly patients with vertebral fractures that do not improve as expected, rather than being done immediately; and the corset was just a thousand-dollar error. Optimal care could not have cost more than a few thousand dollars. But every provider got paid for all those extra activities, and they would have made money on the vertebroplasty, too.

A colleague once pointed out: “If we paid religious leaders by the prayer, there would be a lot of prayers said.” In Medicare, we still mostly pay “by the service,” so we get a lot of services, including those with very low expected value, and sometimes even harmful ones. One might think that managed care would vigorously go after these opportunities for savings; but until recently, their efforts in this regard have been thin, in part because few clinicians recognized frail elders as any different from younger adults, in whom testing for infrequent etiologies might be justifiable. With geriatricized medical care (Core Component #3, page 47), the elderly person and family are clearly placed in a position of control, and the care plan (Core Component #2, page 23) reflects their priorities and goals. In our current climate of overuse, good decision-making and care planning will reliably reduce medical costs.

To create a full prototype that other communities could look to for guidance, we analyzed how the MediCaring Community model would work in these four diverse geographic areas: Akron, Ohio; a section of Queens, New York; Milwaukie, Oregon (a suburb of Portland); and Williamsburg, Virginia. We researched the range of reported savings in various initiatives and conferred with clinical leaders in each site. From that process, the teams came to consensus as to how much change they could confidently expect to make within 18 months, which we assumed as the start-up period. The major literature sources and the teams’ consensus estimates are given in Table 6.1. These estimates, along with Table 6.1, have been published in Milbank Quarterly.[163]

Table 6.1: Estimates of Effects of Implementing MediCaring in Four Communities

Service	Estimates From Published Research and Pilots		Teams’ Consensus on Predicted Changes within 18 months
	Estimate	Notes
Inpatient Hospital	-29% to‑66%	2009; NY Independence at Home Act[164]	-25%
	-8% to ‑33%	2012; demonstrations to cut risky hospitalization[165]
	-10%	2006; resource use among elders receiving acute care[166]
	-17%	2011; hospital admissions/savings with INTERACT II[167]
	-18%	2005; two-year GRACE implementation[168]
	-36%	2011; FFS readmits with Care Transitions Intervention[169]
	-61%	2012; readmissions after psychosocial counseling[170]
Outpatient Hospital	7%	2005; two-year GRACE implementation[171]	10%
Emergency Services	-10%	2005; two-year GRACE implementation[172]	-25%
	-35% to ‑59%	2009; NY Independence at Home Act[173]
Primary Care	-1%	2005; two-year GRACE implementation[174]	30%
Professional Specialty Care	-36%	2005; two-year GRACE implementation[175]	-15%
	-53%	2005; hospital-at-home model[176]
SNF	-15%	2012; cost-containing care transition strategies[177]	-20%
Home Health	20%	2012; cost-containing care transition strategies[178]	10%
Hospice	67%	2013; advance care planning and quality outcomes[179]	10%
Ambulance		Estimates obtained through local survey	-25%
Transporta­tion		Estimates obtained through local survey	100%
Medicaid-covered long-term care	-28%	2013; PACE versus Medicare FFS expenditure[180]	-5%
	-71%	2012; Diversion of nursing home eligible patients to home and community services[181]
	-72%	2006; interventions for Alzheimer’s patients’ spouses[182]

---

We accounted for participants dying or moving away and we used the local estimates of the likely proportion of frail elderly people who would participate. We estimated recruitment, training, equipment, and marketing costs at each site and the development of performance standards and quality metrics jointly. The model was conservative in estimating that the improvements would only be 50% effective in the first year, 90% in the second, and 100% in the third. Overall savings in the steady state would take 20% from baseline medical costs and 5% from institutional long-term care costs. Hospital and skilled nursing facility costs would decrease, but primary care and home care costs would increase. The sites used their local personnel, training, and information technology costs, as well as setting out their best available strategy for implementation, including enrollment.

When we put it all together, the MediCaring implementation yielded a 91% return on investment (ROI) during the startup period and a 249% ROI in the steady state, with the program then projected to be able to sustain operations derived from savings. The total savings over the first three years, as confirmed by an independent actuarial assessment by Ernst & Young, was projected at $57 million, including the costs of developing new methods for quality measurement, implementing program evaluation and quality improvement, setting clinical standards, supporting community organizing, and cross-community collaboration, as well as the on-site start-up costs. After the first three years, the four communities together would generate annual savings of $31 million to use in expanding supportive services, monitoring and managing their system, paying back the startup funds, and sharing with Medicare. A brief summary of the yield at each site is shown in Table 6.2.

Table 6.2: Steady State Estimated Financial Performance for Four MediCaring Communities

	Participants in Year 3	Net Savings in Year 3	PMPM* savings, Year 3
Akron, OH	2,924	$17million	$328
Williamsburg, VA	1,396	$5 million	$269
Milwaukie, OR	828	$3 million	$291
Queens, NY	1,483	$6 million	$537
Total	6,631	$31 million

*Per Member Per Month

---

Clearly, substantial savings are possible. Even in a very low cost Medicare market like Milwaukie, Oregon, the ROI was positive in the second year and the savings were substantial. In a very high-cost area like Queens, savings per person are higher and ROI becomes positive very quickly. The timetable and extent of savings depends on elements of the context and the intervention such as the historical spending pattern, the speed of implementing better geriatric medical care, and the community’s trust in the modified care arrangements to encourage participation. However, with our conservative estimates of effectiveness and enrollment, the improved delivery system in a MediCaring Community has substantial opportunity to create Medicare savings.

Recently, the opportunities for savings have been illuminated by the success of companies serving as conveners for bundled payments, such as naviHealth and Remedy Partners. In bundling post-hospital care, naviHealth has posted a report that says that their work saves nearly half of the 10% of Medicare payments that fall into the 90 days after hospitalization.[183],[184] The company and the sponsor of the bundle thus earn almost 5% of the entire Medicare budget for their proportion of the beneficiary population, mainly by setting the expectations of patients and families for shorter lengths of stay in skilled nursing or rehabilitation facilities after hospitalization and by channeling some patients into lower cost post-hospital settings. These efficiencies do not worsen health outcomes, and in some cases they actually avoid harms. These strategies do put more burdens on family caregivers or on hired helpers for personal care, but Medicare does not cover these services.

The savings are clearly possible and substantial. Returning these savings to support community services would substantially reduce waiting lists and address gaps in support for frail persons living in the community. The next question is whether an integrated approach with enhanced LTSS would provide further savings, and then whether a MediCaring Community can enable a part of the Medicare savings to be reinvested to optimize eldercare in the community.

---

[160] (Buchbinder, et al. 2009)

[161] (Kallmes, et al. 2009)

[162] (Trout, Kallmes and Kaufmann 2006)

[163] (Bernhardt, et al. 2016)

[164] (New York Academy of Medicine 2009)

[165] (Brown, et al. 2012)

[166] (Jayadevappa, et al. 2006)

[167] (J. G. Ouslander, et al. 2011)

[168] (Counsell, et al. 2009)

[169] (Voss, et al. 2011)

[170] (Watkins, Hall and Kring 2012)

[171] (Counsell, et al. 2009)

[172] (Counsell, et al. 2009)

[173] (New York Academy of Medicine 2009)

[174] (Counsell, et al. 2009)

[175] (Counsell, et al. 2009)

[176] (Leff, Burton, et al. 2005)

[177] (Dobson, et al. 2012)

[178] (Dobson, et al. 2012)

[179] (Bischoff, et al. 2013)

[180] (Wieland, et al. 2013)

[181] (State of Florida Department of Elder Affairs 2012)

[182] (Mittelman, et al. 2006)

[183] (Linehan and Coberly 2015)

[184] (Scully 2014)