Multiple studies have confirmed the existence of a population of patients—in the U.S. Department of Veterans Affairs (VA) health care system and across all health care systems—with complex health care needs who are at elevated risk for further morbidity, hospitalization, or mortality in the near term. This small group of patients uses a disproportionate amount of care resources. Contributing to the challenges of care planning for this population is that individuals are highly heterogeneous, both at any given time and over time. No two patients present the same set of symptoms, diagnoses, and challenges related to social determinants of health (SDOH). Furthermore, high-risk status is fluctuating in nature: Among an identified high-risk population, some will die within two years, some will remain at chronically high risk, and some (as many as two-thirds) will achieve reduced risk. In addition, approaches for identifying the appropriate patients for targeted interventions, such as those that use electronic health records, might be insufficient. Thus, attempts to define the meaning of high-quality care for high-risk patients and to establish its evidence base have been frustrating.
In late 2020 and early 2021, a team of RAND researchers conducted a study for the VA using a virtual RAND-University of California, Los Angeles, appropriateness panel to identify health care assessment priorities and associated quality-of-care measures or measure concepts for patients identified as being at high risk. When we refer to a measure, we mean the specification of how, when, and with what population an evidence-based care process guideline should be implemented such that it can be used to assess care quality at the individual and system levels. Measure concepts are those guidelines or recommendations that have not been developed into measures. The panel comprised a group of experts in the care of patients with complex health care needs, including primary care physicians, nurse practitioners, social workers, and health services administrators—most with experience working in the VA health care system—and several VA patients. The panelists were asked to assess a proposed set of evidence-based potential measure concepts and a group of existing VA measures currently in use that address many of the same measure concepts. The existing measures specified the target patients and other details, such as recommended frequency or tools.
The final ratings showed 22 measure concepts that the panelists regarded as high priority and with strong agreement. With these 22 concepts that might merit consideration for inclusion in a measure dashboard for high-risk patients in primary care clinics, the research team first compared the measure concepts to existing VA measures to identify any that had identical or very similar existing VA measures already in use, because, in such cases, no additional development or validation work would be required to incorporate them into the dashboard. It would instead be a matter of repurposing existing measures for this new use. This evaluation found that 16 highly rated measure concepts had existing VA measures that could be repurposed. However, the remaining six highly rated measure concepts had no comparable VA measure and, therefore, are candidates for inclusion in a measure dashboard if evidence can be found that links the use of such measures to better care outcomes. Such evidence would hopefully be in the form of existing evidence-based guidelines and measures.
The primary focus of this project was to determine whether there are existing, operational quality measures for two of the remaining six measure concepts, selected by the VA, and whether these measures have been validated for use in primary care settings by primary care clinicians. The two concepts are as follows:
- identification of a patient's need for and/or source of social and caregiver support and referral to social services if the patient has been identified as high risk and with social need (this was considered two separate measure concepts in the panel project)
- assessment of cognitive impairment (CI).
We conducted two scoping reviews to identify existing measures, guidelines, and tools for (1) assessing social support and the need for caregiver support and referral to social services and (2) screening for CI among patients with complex health care needs in primary care settings and the need for referral for further testing. The two reviews followed published guidance for conducting and reporting on scoping reviews.
- We identified no existing measures for assessing existing social or caregiver support or the need for referral to social services in primary care settings.
- One identified guideline provided a decision support tool for assessing the need for caregiver support. Information on validation of this tool was not found.
- Social support, as conceived by the tools and other resources identified for the current study, encompasses at least two concepts: help when needed with performing activities of daily living (ADLs) and presence of emotional support (i.e., the absence of loneliness or social isolation).
- Many tools exist for assessment of ADLs and SDOH in primary care settings, and some of these tools include items to assess social support, need for further caregiver support, and need for referral. None of the tools or items was adapted for patients with complex care needs or validated using health or other important outcomes.
- We identified several measures and evidence-based guidelines for screening patients for CI in primary care settings, although routine screening of patients who do not present with evidence of CI remains controversial.
- Numerous brief screening tools exist for assessment of age-related CI, and several tools exist for screening for CI associated with chronic conditions. Most of these tools have been validated against specialty assessment, and some have been compared in the same populations.
- No one screening tool appears to be ideal, but some candidates exist. Nevertheless, several challenges must be overcome for all CI screening. Chief among these is the lack of applicability of screening tools to patients with low education and low literacy and the lack of validation among diverse populations.
It is beyond the scope of this study to make research or practice recommendations. Nevertheless, the literature suggests the following next steps:
- Social support and caregiver screening of patients with complex care needs for referral appear feasible and necessary, but it will be important to identify brief screening tools that can be used with all patients or to identify criteria for which patients to assess further.
- Social support screening will need to undergo validation in populations of patients with complex care needs.
- Cognitive screening of patients with complex care needs in the primary care setting appears feasible, but patient factors that can affect performance and choice of screening test must be considered carefully.
- Risks, benefits, and cost-effectiveness of screening complex patients (versus screening only patients who demonstrate some signs of need for social support or present symptoms suggestive of possible CI) must be considered.