Avoiding Therapeutic Nihilism from Complex Geriatric Intervention “Negative” Trials: <scp>STRIDE</scp> Lessons

Abstract

“It ain't so much what you don't know that gets you into trouble, as what you know that ain't so!,” Will Rogers 1930s (?) In the early days of the Journal of the American Geriatrics Society, the multifaceted complexity of aging was recognized in conjunction with a hypothesis that “relatively radical, vigorous, usually expensive, extensive therapeutic intervention may result in noteworthy rehabilitation of markedly debilitated aging and aged subjects.”1 Two years later, Dr Joseph Sheldon observed that scant research existed around the problem of falls, simplistically attributed to just four causes.2 Fast-forward 65 years and falls remain a monumental problem for many older adults, linked to 32,000 deaths and 950,000 hospitalizations or transfers as over 27% of the U.S. population aged 65 years and older falls every year.3 A multistakeholder panel that included the American Geriatrics Society in 2010 provided clinical practice guidelines to evaluate fall risk and initiate efforts to reduce associated injuries.4 The U.S. Preventive Services Task Force (USPSTF) recently recommended select use of multifactorial interventions to reduce falls in at-risk community-dwelling older adults.5 Despite these recommendations, falls prevention efforts often seemed inadequate across healthcare settings, possibly attributable to an imperfect fit between research settings and real-world clinical care.6 The Patient-Centered Outcomes Research Institute and the National Institute on Aging funded the Strategies to Reduce Injuries and Develop Confidence in Elders (STRIDE) in 2014 to begin overcoming some of these practical barriers between research settings and everyday clinical practice in reducing fall morbidity via risk-assessment guided individualized care plans informed by the Chronic Care Model to balance evidentiary rigor with local adaptability.7 STRIDE included 86 primary care practices from 10 healthcare systems recruiting community-living, cognitively intact individuals 70 years or older (lowered from the original design ≥75 years due to recruitment issues) identified as increased risk for falls if they had experienced a fall-related injury in the last year or had two or more falls in the prior year or were afraid of falling. Unfortunately, this well-funded, multicenter, pragmatic, cluster-randomized trial did not demonstrate a reduction in first serious fall injury.8 So how do clinicians, researchers, and guideline developers proceed without slipping into a counterproductive mindset of therapeutic nihilism in aging research? Although randomized controlled trials are perceived as the optimal method to understand the potential benefits or harms of an intervention by theoretically distributing confounders equally across treatment arms, their focus is on quantifying efficacy so research designs favor ideal conditions often inconsistent with actual clinical practice, which require effectiveness studies outside the research setting. Understanding effectiveness in real-world settings necessitates minimizing exclusion criteria and application of interventions by nonresearch personnel amidst the chaos of daily schedules, competing clinical priorities, nonadherent patients, and personal financial constraints. The reward for effectiveness research is a more confident understanding of external validity outside the ivory tower, whereas the disadvantage is lesser ability to understand whether the intervention might have worked in an ideal setting.9 Pragmatic trials ideally seek to enroll every eligible patient by waiving informed consent when possible, recruiting investigators across rural and nonacademic hospital settings, and delivering the intervention in normal practice using routine staff and equipment.10 In contrast to the simplistic views of 1955, falls often represent a complex web of aging physiology effecting gait, balance, sensory perception, cognition, and reflexes overlying medication adverse effects and home hazards. This underlies the USPSTF recommendation for individualized multifactorial interventions rather than far more simple interventions, such as eyeglasses or medication deprescribing. STRIDE is hardly the first complex intervention in geriatrics to render unexpectedly negative results (Table 1).11-13 In conjunction with STRIDE, these “negative” studies provide important lessons for future geriatric researchers testing solutions to complex health issues with multicomponent interventions. Elucidating benefit probably requires a formal evaluation of the multiple layers of potential confounders at the level of the patient and caregiver, healthcare providers responsible for administering the intervention, and contextual factors of the environment in which patients/providers interact. Patient complexities unreported in STRIDE or any of the studies in Table 1 include frailty, health literacy, social isolation, sensory deficits, or financial constraints—any or all of which could be associated with adherence to fall risk reduction strategies. Caregiver issues are similar and could also include psychosocial strain. Healthcare provider confounders could include baseline communication skills, inherent expertise and interest in geriatrics, and available time on task—each of which could influence the fidelity of the intervention during the crucial moments of interface with patients. The setting in which the intervention is delivered might be affected by organizational capacity to adapt, differential access to follow-up care, and preexisting commitment to or institutional incentives to optimize older adult outcomes. Any or all of these confounders could skew results and mask subsets of patients likely to benefit (or be harmed) from well-intentioned, complex interventions. STRIDE investigators sent 31,872 recruitment packets, completed 18,571 telephone interviews, and enrolled 5,451 patients.8 However, obtaining patient follow-up required significant efforts, and patient adherence to recommended fall prevention strategies was not monitored. Incorporating implementation science principles into future complex geriatric intervention trials might overcome some of these barriers. Standards for Reporting Implementation Studies (StaRI) did not exist when STRIDE launched, but now provide an infrastructure to distinguish intervention effectiveness from implementation strategy impact.14 Adhering to StaRI would include explicit descriptions of the context in which the intervention is implemented, including personnel resources and social, economic, policy, and organizational barriers or facilitators. In addition, StaRI mandates a description of the theoretical framework underlying behavior adaptation for patients and clinicians. The Chronic Care Model is a form of healthcare delivery, but not an explicit behavior change model, of which dozens exist.15 Beyond theoretical frameworks, StaRI methods also measure and report fidelity of the implementation strategy and delivery of core components of the intervention, as well as a more formal assessment of external validity and potential harms or unintended effects.16 Patient engagement is important. STRIDE investigators employed motivational interviewing to promote patient buy-in. The science of shared decision-making is rapidly evolving to inform more meaningful patient-clinician information exchange and theoretically empower patient buy-in.17 Shared decision-making builds on the patient philosophy of “no decision about me without me,” often requiring a formal assessment of decision quality with varying tools and approaches for different scenarios. The setting for falls prevention could also make a difference. For example, patients engaged during the teachable moment of an emergency department visit for a fall might be more receptive and adaptive to evidence-based interventions using shared decision-making.18 The STRIDE measures used to assess fall risk, although straightforward and reproducible, might also merit reevaluation. Dozens of fall-risk instruments exist across healthcare settings. Identifying feasible and accurate risk assessment strategies occurs locally, but some efforts to achieve harmonization of measures across specialties could improve comparability.19 The same harmonization of geriatric measures could apply to evaluations for cognitive dysfunction and frailty.20, 21 Cognitive impairment also occurs on a spectrum, and ethical approaches to recruit some of these individuals into trials now exist, yet STRIDE excluded those with an abnormal Six Item Screener.20, 22 Finally, the trial outcomes and effect size are worth contemplating. STRIDE demonstrated a reduction in falls but not fall injuries. If fall-related healthcare expenses exceed $50 billion per year, an 8% reduction in falls might seem clinically trivial but potentially equate to $4 billion savings. The path forward depicted in Figure 1 for transdisciplinary clinicians, researchers, and patient advocates requires disruptive innovation that incorporates these observations. However, in the meantime, astute opinion leaders must be prepared for a potential nihilistic backlash. For example, guidelines and accreditation criteria catalyzed geriatric quality improvement in emergency medicine and surgery over the last 5 years.23, 24 Yet, guidelines are frequently criticized and mistrusted by clinicians, partially because they are based on consensus rather than high-quality evidence of benefit.25, 26 Rather than an instinctive approach to do something actionable when confronted with a medical problem like falls, some philosophize that deliberate clinical inertia (“don't just do something, stand there”) is a wiser and safer clinical approach while awaiting definitive research evidence verifying an intervention's benefits.27, 28 The rapid-learning health system conceptual strategy approach guides us to use the evidence from what works and what does not work to influence continuous improvement.29 One truth is that inaction guarantees that falls will continue to occur in an aging world as they did in 1954. STRIDE proves that reducing injurious falls is exceedingly challenging, which is why the interventions are complex and multifactorial. Detailed attention to unmeasured confounders through the lens of implementation science, while employing explicit shared decision-making and harmonized measures, may yet guide us to reproducible falls prevention. C.R.C. and M.L.M. are faculty for the Geriatric Emergency Department Collaborative. C.R.C. serves on the American College of Emergency Physician's Geriatric Emergency Department Accreditation Board of Governors. C.R.C. also coauthored the Standards for Reporting Implementation Studies reporting guidelines, served as colead of the 2016 Academic Emergency Medicine Consensus Conference on Shared Decision Making, and served as co–principal investigator of the Grants for Early Medical/Surgical Specialists' Transition to Aging Research (GEMSTAR) U13 conferences. Christopher R. Carpenter and Michael L. Malone are the sole contributors to the concept, design, and preparation of this manuscript. The authors acknowledge Kai Choummanivong's artistic contributions in creating Figure 1. None.

Topics & Concepts