Balancing evidence-based care with patient-centered individualized care

Abstract

Weak evidence, when manifested in clinical guidelines, can translate into biased vascular care. In vascular surgery, we have few randomized controlled trials with appropriate representation of females and persons of color, so generalizability of trial results can be problematic. Physicians are required to balance evidenced-based care (which is only as good as the underlying evidence) with personalized treatment recommendations that are often based on demographics, social circumstances, and/or existing therapeutic relationships. Biases, whether implicit or explicit, have an oversized effect on treatment decisions, and patient outcomes. In this commentary, we propose three principles to strengthen the vascular surgery evidence foundation and patient-centered decision-making going forward: (1) generating evidence designed for individualized care, (2) constructing clinical guidelines that are context specific and complexity aware, and (3) strengthening the training and support for surgeons to deliver patient-centered individualized care.

In an ideal world, medical decision-making would be based on high-quality, published evidence and generated to support personalized care based on each individual’s unique health status, disease state, and preferences. This idealized world, of course, is disparate from the one vascular surgeons often encounter. Indeed, we are often faced with making decisions and guiding patients through the shared decision-making process with weak evidence and little insight into how unique patient factors may affect outcomes. Following evidence-based guidelines to reduce random variation in care is an important step toward providing unbiased, high-level care, but our attempts are complicated by the fact that evidence is often based on retrospective studies and not generalizable to females and persons of color owing to a lack of representation in the cohorts studied. Consequently, even in the setting of truly patient-centered shared decision-making, weak evidence leaves room for unwarranted care variation. Put otherwise, weak evidence can lead to bias. In vascular surgery, biases, whether implicit or explicit, have an oversized effect on treatment decisions and patient outcomes. Improving our evidence base and creating data-supported clinical algorithms may lead to more equitable and even improved treatment outcomes.

GUIDELINES ARE ONLY AS GOOD AS THE EVIDENCE

In medicine, evidence-based clinical care is a combination of three main components: (1) the best available scientific information, (2) clinical judgment, and (3) patient preference.^1,2 All three elements are necessary to provide the best care possible to patients, but the first component is what is used to develop clinical practice guidelines. When evidence is weak, guideline recommendations are weak. For example, just a decade ago, in the top five general surgery journals, the sex of trial participants was only stated approximately 80% of the time, and, of the studies that reported sex, only one-third statistically analyzed their results stratified by gender.³ Further, not even one of these studies had equal representation of females compared with males.³ These omissions may not necessarily indicate bad intentions or poor study execution, but rather the realities of study design. For a study to be powered to detect a meaningful effect and achieve a desired level of precision, hundreds or thousands of participants may be needed. Further, that number of study participants is estimated to be the number of participants sufficient to estimate population-based results and is not powered to detect effects for unique individuals or subgroups. Although subgroup analyses can be performed to evaluate treatment efficacy in particular groups, without proper study planning and careful participant accrual monitoring, recruiting enough patients across sex, race, ethnicity, age, and other demographic or comorbidity profiles to produce reliable subgroup estimates is not consistently satisfied.

In vascular surgery, we should be proud that in the 97 arterial trials published since 2008, all of them reported sex distribution in their Table 1 with demographics and comorbidities. Still, the under-representation of females is stark.⁴ Accounting for the sex differences in the prevalence of the disease, females have been consistently under-represented. Even with a heightened awareness of the importance of diversity, equity, and inclusion, there has not been any improvement in the proportion of female trial participants over time—only 16% of the 69 trials in peripheral artery disease and 18% of the 16 trials in thoracic and abdominal aortic aneurysm (AAA) disease had appropriate representation by sex.⁴ Representation in venous interventional trials are no better. Although there are small variations in incidence related to age and hormone status, over the lifespan, deep venous thrombosis is equally prevalent in females and males.⁵ A Cochrane review of thrombolysis vs anticoagulation for the treatment of acute deep venous thrombosis included 19 RCT with 1943 patients, but the review only stated that both males and female subjects were included in the trials, without any mention of the proportions or sex-stratified results, despite known differences in thrombosis and bleeding risks by sex.⁶ The Acute Venous Thrombosis: Thrombus Removal With Adjunctive Catheter-Directed Thrombolysis (ATTRACT) trial was the largest and most contemporary trial included in that review; 53% of the 692 subjects were male.⁷ Another Cochrane review summarizes four randomized controlled trials (RCTs) studying the effectiveness of surgery for deep venous insufficiency; only two of the four reported sex, and there were 33% more males than females studied in those two valvuloplasty trials.^8–10

Although clinical trials are considered the gold standard for producing guideline-worthy data, in many vascular guidelines, data from cohort studies and retrospective reviews of clinical registries and electronic health records (EHRs) are also included. Evidence from EHRs and registries like the Society for Vascular Surgery’s Vascular Quality Initiative is often more representative of actual day-to-day clinical practice and is complementary to RCT data, which is generated under highly controlled circumstances that do not always simulate real life.¹¹ Although to a lesser degree compared with RCTs, registry data are also plagued by under-representation of females and racial/ethnic minorities, because these groups are almost uniformly underdiagnosed and undertreated; therefore, they do not get the procedures necessary to be included in a registry.

Although registry and EHR data sources experience problems with missing data and other biases, there are expanding methods to account for these shortcomings and we are just scratching the surface of the potential that these data sources hold.¹² In North America and Europe, these data sources are readily available to clinicians and researchers to investigate clinical questions quickly. Future efforts should focus on training to further streamline charting in EHRs across sites and platforms and additional support should be provided to set up registries that focus on marginalized groups and registries in countries where there is historically little data like those in Asia and Africa. Accomplishing these goals could help us to close the gap in knowledge for some of the under-represented groups discussed previously and lead to an overall better evidence base to support clinical care.

The implication of the under-representation of females in clinical research (including randomized trials, cohort studies, and retrospective reviews) manifests in poor or absent guidelines for female patients. For instance, consider gonadal vein reflux causing varicocele and pelvic venous disorders in males and females, respectively. Despite established diagnostic criteria and minimally invasive treatment options, pelvic venous disorder is often a missed diagnosis that is generally lumped into noncyclical chronic pelvic pain.^13,14 The Agency for Healthcare Research and Quality has identified a “ubiquitous” lack of research in domains of etiology, diagnosis, and treatment related to pelvic pain, including venous disorders, in females.¹⁵ In stark contrast with the male experience with gonadal vein reflux and varicocele, lacking evidence in females with gonadal vein reflux and pelvic venous disorder results in inconsistent terminology and outcome measures, no widely adopted clinical practice guidelines, and no standard insurance reimbursement offered for treatment of the condition. Another prime example of the consequence of female under-representation is that, despite immense advances in the care of AAAs, we do not really know the safety and efficacy of different repair options in females. With this limited evidence base, it is no wonder the US Preventive Services Task Force recommends against screening females without a family history for AAA and cannot make recommendations for females who have a family history or have been smokers, who are potentially at higher risk for AAA.¹⁶

More broadly, important demographic factors, including race, ethnicity, and age, in addition to sex, have been repeatedly linked to differential disease progression and outcomes. Consequently, it is no longer acceptable to assume that the results gathered from primarily White male populations can be applied to women and other minority groups. Yet, when clinical guidelines are based on studies without appropriate representation, generalizing guidelines to the unique patients we encounter in our clinics and operating rooms becomes difficult at best, impossible at worst, and almost certainly results in inconsistent care. Establishing appropriate clinical guidelines across groups provides a consistent place to begin our care plans and makes it easier for us to consider more individualized care beyond those initial steps.

BIASES IN TREATMENT BASED ON THE EVIDENCE BASE AND IN INDIVIDUALIZED EXPERT OPINIONS

Currently, individualized care is based on each unique physician’s experience and shared decision-making with each unique patient. Although the importance of this patient-centered bond and the influence of personal experience should not be overlooked,¹⁷ it is clear that “doctors often fail to diagnose stereotypical ‘male’ conditions in females, and vice versa, until the condition has become dangerous.”¹⁸ For example, symptoms commonly reported by females are considered atypical because males do not typically report them. Our differential diagnoses and treatment algorithms are still predominately based on male data, so, despite growing evidence of sex differences in vascular diseases, we are not trained to consider the possibility that females may have vascular disease. This bias, unfortunately, leads to many delayed diagnoses.¹⁸

In medical school, physicians are taught the incidence and prevalence of diseases in different patient populations with different presenting signs and symptoms. This skill is required to create a differential diagnosis, and ultimately order the correct diagnostic testing to confirm the most likely diagnosis. The skill and agility used to solve diagnostic problems accurately and efficiently are the marks of a master clinician and is often built over years of practicing the art of medicine. However, when the process is supported by systemic cultural biases and incomplete evidence, expert opinion is threatened by bias (Fig 1).

Fig 1.

Shared decision-making based on weak evidence foundation can lead to unwarranted variations in care and ultimately bias.

More troubling, once the diagnosis is made, females are less likely to be treated for peripheral arterial disease and AAA, despite clear limb- and life-saving advantages.¹⁹ This finding is particularly significant for peripheral arterial disease, because the prevalence is the same for men and women.²⁰ Even when treated, females and tend to be older than males at the time of surgery, suggesting a delay in care.²¹ This finding is also true for minority racial/ethnic groups compared with White individuals.²² Another persistent debate is about the efficacy of open vs endovascular procedures in women. Without sex-specific results from appropriately representative randomized control trials, the answer to this question remains unspecified.

TOWARD EQUITABLE, BUT STILL INDIVIDUALIZED, CARE

Weak evidence, when manifested in clinical guidelines, can translate into biased vascular care. As a starting place, we need more prospective studies in all areas of vascular surgery. This finding is especially true because we are often limited by low-quality, retrospective evidence to support guideline recommendations. How we conceptualize and design new studies is an opportunity to shape a more equitable, patient-centered future for vascular medicine. To realize this opportunity, we propose three principles to strengthen the vascular surgery evidence foundation and patient-centered decision-making going forward: (1) generating evidence designed for individualized care, (2) constructing clinical guidelines that are context specific and complexity aware, and (3) strengthening the training and support for surgeons to deliver patient-centered individualized care (Fig 2).

Fig 2.

A stronger evidence foundation can help surgeons balance evidence-based guidelines and individualized care in shared decision-making.

Precision medicine: Evidence for individualized care.

One promising approach to building evidence for individualization is the precision medicine paradigm. Colloquially, precision medicine is defined as matching the right patient, to the right treatment, at the right time, and methodologically, precision medicine approaches range from the development of highly targeted genomics-focused therapies to advanced analytics that learn from data on how to optimally match existing treatments, in sequence, to individuals based on their genes, disease presentation, responses to previous treatments, and lifestyle and environmental factors.^23,24 The benefit of bringing a precision medicine lens to vascular surgery study design is that the precision medicine paradigm embraces and leverages patient heterogeneity. Recruiting clinically and demographically diverse participants is paramount for the application of precision medicine analytics and learning how to best tailor treatments to different types of patients. In the near term, a focus on this type of evidence generation for individualization can help to prioritize representation in study design and recruitment and support more generalizable clinical guidelines. In the future, precision medicine evidence may be used to generate algorithms for sequential, individualized clinical decision-making at the bedside.²⁵

Systems-aligned precision medicine: Context-specific, complexity-aware guidelines.

Physician and patient decision-making are shaped by the system in which health care is delivered and received, including patient preferences, inequities in health care access, exposure to structural and interpersonal sexism and/or racism, and physical and resource constraints. Although data-supported precision medicine has significant potential to offer real-time individualized decision support, evidence-based guidelines will be rendered ineffective at best or unjust at worst if they are infeasible, unpractical, or inaccessible to real patients in the real world. As with any evidence generation paradigm, evidence generated using precision medicine analytics alone is not a panacea for systemic sexism, racism, and other bias in health care. Indeed, precision medicine evidence generated from data that systematically excludes minority and underprivileged groups will not be useful in the treatment of those patients. And, even if the evidence basis of precision medicine solutions are based on appropriately representative patient cohorts, if those solutions are not accessible or made available to patients equitably, for example, owing to costs, they will only exacerbate existing structural inequalities.^26,27

A systems-aligned precision medicine approach advocates for grounding evidence for individualized care in the system in which that care will be delivered.²⁸ This means defining the system, the set of embedded elements that interact to influence outcomes, and engaging patients and stakeholders to inform every step of the evidence generation and implementation process. Necessarily, a systems-aligned approach will require a mixed methods approach and may include strategies from participatory systems science and community-based research to build trust, identify needs, and understand the types of solutions that work best for different groups given their historical relationship to healthcare, societal and cultural norms, and resource constraints. Additional systems-aligned precision medicine strategies may include the use of rigorous evidence generation frameworks such as the Data Equity Framework²⁹ that center equity in data collection and analysis. How this will manifest in vascular surgery research is multifaceted. For example, there is scant literature describing medical uncertainty, health-related quality of life or functional status, or shared decision-making about vascular conditions and treatments from a patient’s perspective. Capturing these outcomes and bringing them to conversations with patients, in addition to standard surgical and technical success outcomes, will be part of centering patients in decision-making.

Supporting equitable care: Training for surgeons in patient-centered individualized care.

Ultimately, medical decision-making is made at the bedside and mediated through clinicians who are juggling their caseload, administrative duties, and pressure to diagnose and treat diseases as quickly as possible. Assisting patients through difficult decision-making processes, even with ideal evidence designed for individualization, is hard, much less with imperfect guidelines made for the average patient and not necessarily the unique patients we every day with their own values and goals. Medical schools and postgraduate programs can, and should, provide formal teaching in truly individuated, patient-centered communication that can be performed in the current medical setting as opposed to an idealized setting without time constraints. Forming a partnership with each patient and relating the technical information either from evidence-based guidelines or expert opinion in a way that fits with patients’ values and expectations is key to delivering equal and equitable care.

CONCLUSIONS

Balancing the recommendations of evidence-based guidelines, expert opinion and experience, and patient values and goals is both the art and the challenge of individualized patient care. When evidence is weak, its void leaves space for biased individualized decision-making. Minimizing this bias is one way to truly center patients in care decisions. We believe that, going forward, this process entails following the principles of building better evidence through paradigms that are designed to generate evidence for individualized care, contextualizing evidence to be pragmatic and feasible for the real patients that we see in our clinics and operating rooms, and supporting physicians who facilitate and mediate the shared decision-making process.

As a field, we have an opportunity to improve the way we provide vascular care. Balancing evidence-based and individualized care may not ever be easy, but we most certainly can build for ourselves the evidence to make it less hard and more robust to potential bias. Ultimately, that will allow us to provide evidence-based guideline care that is also truly patient-centered individualized care.