ORLANDO – Accumulating experience is showing the benefits of various models of care for cancer survivors in terms of health care use and costs, while also suggesting that some provide higher-quality care than others, according to a pair of studies reported at a symposium on quality care sponsored by the American Society of Clinical Oncology.

Initiative for breast cancer survivors

“In 2011, Cancer Care Ontario did a quick environmental scan of our 14 regional cancer centers and found that the transition of breast cancer survivors from oncologists to primary care was very variable, and that centers often didn’t transition patients very frequently,” said Nicole Mittmann, PhD, first author on one of the studies, chief research officer for Cancer Care Ontario, and an investigator at Sunnybrook Research Institute, Toronto.

The advisory organization therefore implemented the Well Follow-Up Care Initiative to facilitate appropriate transition of breast cancer survivors. Each regional center was given a $100,000 incentive to roll out a model of the initiative.

Dr. Mittmann and her coinvestigators used provincial administrative databases to compare health care use and associated costs between 2,324 breast cancer survivors who were transitioned with the initiative and 2,324 propensity-matched control survivors who were not. The survivors were about 5 years out from their breast cancer diagnosis at baseline and had median follow-up of 2 years.

Study results reported at the symposium showed that the mean annual total cost of care per patient paid for by the provincial health ministry was $6,575 for the transitioned group and $10,832 for the nontransitioned group, a difference of $4,257 (39%). The main drivers were reduced costs of long-term care and cancer clinic visits.

Findings were similar for median annual costs, which amounted to $2,261 for the transitioned group and $2,903 for the control group, a difference of $638.

Compared with the nontransitioned group, the transitioned group had significantly fewer annual visits to medical oncologists (0.39 vs. 1.29) and radiation oncologists (0.16 vs. 0.36), while visits to general or family practitioners were statistically indistinguishable (7.35 and 7.91), Dr. Mittmann reported. There was also a trend toward fewer emergency department visits.

The transitioned group had fewer bone scans, CT and MRI scans, and radiographs annually, but differences were not significant.

Reassuringly, Dr. Mittmann said, survivors who were transitioned did not fare worse than their nontransitioned counterparts in overall survival; if anything, they tended to live longer. “We think that because the individual cancer centers enrolled patients that they thought were very well that this is a very well and highly selected and maybe a biased group,” Dr. Mittmann acknowledged. “But we certainly see that they are not doing worse than the control group.”

“About $1.4 million was distributed to the cancer centers” for the initiative, she noted. “That generated a savings for the health system of $1.5 million, if you are looking at median costs, to $9.9 million, if you are looking at mean costs.

“The transition of appropriate breast cancer survivors to the community appears to be safe and effective outside of a clinical trial, at least based on this particular retrospective analysis using databases,” she said. “The overall costs are not increased, and they may actually be decreased based on our data, and certainly these results will inform policy.”

The investigators plan several next steps, such as encouraging senior leadership at Cancer Care Ontario and the Ministry of Health to endorse the findings, according to Dr. Mittmann. In addition, “[we plan to] engage with both oncology and primary care leadership and think about how we can potentially roll out a program like this, and develop tools, whether those are letters or information packages, and education, to … appropriately transition individuals.”

Considerations in interpreting the study’s findings include the quality of the matching of survivors, according to invited discussant Monika K. Krzyzanowska, MD, a medical oncologist at Princess Margaret Cancer Centre, an associate professor at the University of Toronto, and a clinical lead of Quality Care and Access, Systemic Treatment Program, at Cancer Care Ontario. “The quality of that match depends on what’s in the model, so there could be potential for residual confounding, and administrative data may not have all of the elements that you would need to get a perfect match.”

Additional considerations include costs not covered by the payer, impact of the initiative on delivery of guideline-recommended care and patient and provider satisfaction, generalizability of the findings, and long-term outcomes.

“This is a proof of concept, certainly, that transition of low-risk cancer survivors to primary care is feasible and potentially economically attractive,” Dr. Krzyzanowska concluded. “It would be useful to have a formal evaluation of effectiveness that would inform a comprehensive value assessment. And we do have data from a randomized trial about the safety of this particular approach, but it would be nice to see that following implementation in real practices, those safety considerations played out the same way.”

Comparison of survivorship care models

Two-thirds of the large and growing population of cancer survivors are at least 5 years out from diagnosis, stimulating considerable discussion in the oncology community about how to best address their needs, according to Sarah Raskin, PhD, senior author on the second study and a research scientist at the Institute for Patient-Centered Initiatives and Health Equity at George Washington University Cancer Center, Washington.

“Yet, for a lack of cancer survivorship–specific guidelines from research or practice, cancer centers are increasingly developing survivorship care in a variety of ways, many of which are ad hoc or unproven as yet,” she said.

Dr. Raskin and her colleagues compared three emerging models of survivorship care: a specialized consultative model and a specialized longitudinal model – whereby patients have a single or multiple formalized survivorship visits, respectively, with care typically led by an oncology nurse-practitioner – and an oncology-embedded model – whereby survivorship is addressed as a part of ongoing oncology follow-up care, typically by the oncologist.

The investigators worked with survivors to develop the Patient-Prioritized Measure of High-Quality Survivorship Care, a 46-question scale assessing nine components of survivorship care that capture the health care priorities and needs that matter most to patients. Each component is rated on a scale from 0 (not at all met) to 1 (somewhat met) to 2 (definitely met).

Analyses were based on responses of 827 survivors of breast, colorectal, and prostate cancer who received care at 28 U.S. institutions using one of the above models and who were surveyed by telephone about the care received 1 week after their initial survivorship visit.

Results showed that survivors cared for under the three models differed significantly with respect to scores for seven of the nine components of quality of care, Dr. Raskin reported. The exceptions were practical life support, where the mean score was about 0.6-0.8 across the board, and having a medical home, where the mean score was about 1.8-1.9 across the board.

The specialized consult model of care had the highest scores for mental health and social support, information and resources, and supportive and prepared clinicians. The specialized longitudinal model of care had the highest scores for empowered and engaged patients, open patient-clinician communication, care coordination and transitions, and access to full spectrum of care. The oncology-embedded model had the lowest scores. Analysis of the tool’s 46 individual questions showed that patients cared for at institutions using the oncology-embedded model were significantly less likely than were counterparts cared for at institutions using the specialized models to report that the institution performed various activities such as offering a treatment summary, inquiring about the patient’s biggest worries or problems, and explaining the reasons why tests were needed (P less than .05 for each).

For some metrics, the overall proportion reporting that an activity was performed was low, regardless of the model being used. For example, only 48% of all patients reported being helped to set goals or make short-term plans to manage follow-up care and improve health, merely 24% reported being provided emotional and social support to deal with changes in relationships, and just 19% reported being referred to special providers for other medical problems.

“Overall, all three models are performing highly in terms of providing survivors with a medical home and communicating with patients. However, all three are performing quite low in terms of providing mental health and social support, as well as practical life support,” said Dr. Raskin.

“By model, we see that the embedded ongoing care model is significantly underperforming compared with both specialized models on seven of nine components, and we have some hypotheses from our early work with [Commission on Cancer]–accredited centers to explain this,” she added. “Embedded survivorship models have a lot of variability – many are high performers but others are low performers as compared with specialized programs. Embedded survivorship care models are typically led by the treating oncologist, who historically has focused on treating sick patients and less so on providing social supports for follow-up of well patients or ‘well-er’ patients. At the same time, specialized models focus predominantly on survivorship care and providing services and referrals for survivors, which may explain their high scores.

“We know that the higher quality of care measures presented here do not necessarily translate to better patient outcomes, and that’s actually going to be the next phase of our analysis,” she concluded.

The study sample may have had some selection bias, and it is unclear how well validated the tool was, according to Dr. Krzyzanowska, the discussant. Another issue was its assessment of quality of care at only a single time point.

Nonetheless, the findings show “that measuring quality of survivorship care from a patient perspective is feasible and valuable. We have already heard about [need for] survivorship plans in survivorship care, so certainly the work that was just presented is extremely important to help to fill some of these gaps,” she said.

“I’m not sure that we yet know what the optimal model of survivorship care is without the information of the other outcomes. Furthermore, there’s different survivor populations and different ways that health care is organized, so perhaps there isn’t really one optimal model, but the model has to fit with the context,” Dr. Krzyzanowska concluded. “That being said … the tool that they have created can be a great tool for existing survivorship care programs to assess and improve the quality of their care.”

Dr. Mittmann and Dr. Raskin had no disclosures to report.

Survivorship care models work, some better than others