Refocusing the Lens: Patient Safety in Ambulatory Chronic Disease Care

Ensuring patient safety has been recognized as a major public health challenge. According to the Institute of Medicine (IOM), in the U.S. as many as 98,000 deaths per year are attributable to preventable adverse events that occur in the hospital setting, with annual costs (lost income, disability, and health care costs) of between $17 billion and $29 billion.¹ Because the patient safety movement originated in and has focused on acute care settings,² less is known about safety outside the hospital setting. A more recent IOM report concluded that the number of errors affecting outpatients is likely to far exceed those from the inpatient setting.³ As a result, the consequences and costs of medical errors in ambulatory settings may be comparable to or even greater than those in the inpatient setting, making ambulatory research and safety promotion all the more pressing.

Ambulatory visits constitute the overwhelming majority of medical care encounters, and with shorter hospital stays, patients cared for at home are sicker than in the past.⁴ Moreover, performing all recommended actions in the limited time of an office visit has become less feasible,⁵ increasing the potential for unsafe practices and errors. To advance ambulatory patient safety, we propose adapting the Chronic Care Model⁶ to frame the research, practice, and policy agenda. We employ clinical vignettes to illustrate the links between the Chronic Care Model and safe provision of ambulatory care and we use this adapted model to propose avenues for future research and intervention. We aim to link safety and self-management conceptually, so that the systems approach to safety can inform chronic disease management and the patient-centered approach to chronic disease management—one that emphasizes the primacy of the patient—can inform the patient safety movement.

Contrasting Hospital and Ambulatory Safety

The terms patient safety, adverse events, and medical errors have all been used to describe morbidity and mortality associated with medical care and treatment. In this article, we use the IOM’s definition of patient safety: “the prevention of harm to patients.”^{3(p. 5)} The IOM further specifies that both errors of commission, such as prescribing a contraindicated medication, and errors of omission, such as failure to perform recommended medication monitoring, can each jeopardize patient safety. We acknowledge that the concepts of patient safety and quality of care overlap substantially; in fact, the IOM “strongly believes that patient safety is indistinguishable from the delivery of quality care.”^{3(p. 5)} However, the IOM definition of patient safety does not explicitly include patients as active participants in their care. Many argue that in the context of ambulatory care of chronic diseases, highlighting the patient’s central role in safe care is imperative.⁷

In acute care settings, patients are under near-constant observation and often receive care as passive recipients. In contrast, ambulatory patients not only must navigate the health care system, they must actively obtain medical care and make critical self-care decisions. Although recent efforts such as The Joint Commission’s Speak Up^™ initiatives (available at http://www.jointcommission.org/PatientSafety/SpeakUp/) have emphasized the need to activate hospitalized patients and inform about their role in promoting safety, this need also extends to ambulatory patients.

We have chosen to focus this discussion on safety for ambulatory populations with chronic diseases for several reasons. First, chronic disease care represents greater than 75% of all medical care.⁸ Enhancing patient safety in chronic disease care has clear public health relevance. Second, with today’s emphasis on aggressive treatment goals for chronic conditions, medication intensification and high-risk medication use routinely occur in ambulatory care, with corresponding reports of adverse events in chronic disease settings ranging from diabetes,^9,10 rheumatic diseases,^11,12 cardiovascular conditions, and anticoagulation^13,14 to organ transplantation.^15,16 Third, chronic disease patients today take far more daily medications than was routine even a decade ago.¹⁷ Fourth, the high incidence of repeated adverse drug events (ADEs) in the same individuals supports the notion that those with chronic disease face especially high risk for safety problems.¹⁸

Despite this, prior work in ambulatory patient safety has focused on one-time ADEs^19,20 and medical errors.^21–23 These studies have relied on Donabedian’s conceptual framework of structure-process-outcome.²⁴ This causal framework lends itself to understanding discrete events, such as post–hospital discharge errors,²⁵ adverse events leading to emergency department visits,²⁶ or risk-management²⁷ and malpractice claims.²⁸ Such an approach does not adequately address the complexities of chronic disease care in the ambulatory care milieu.

A Conceptual Model for Ambulatory Safety

The field of ambulatory safety research lacks a well-established conceptual model that takes into account the complex, longitudinal nature of chronic disease care. In this article, we extend Wagner’s Chronic Care Model⁶ using theoretical work in patient safety²⁹ to develop a model for chronic disease safety, shown in Figure 1 (page 379). We propose adapting the Chronic Care Model because we believe that for patients with chronic diseases, multiple contributing factors interact in complex, often reciprocal ways, to produce a safe or unsafe state.

Figure 1.

The ecological model for ambulatory patient safety in chronic disease represents an extension of the Chronic Care Model (Wagner E.H.: Chronic disease management: What will it take to improve care for chronic illness? Eff Clin Pract 1:2–4, Aug.–Sep. 1998.). IT, information technology.

To describe the adaptation of this well-established model to ambulatory patient safety, we first discuss the community and health system. Second, we focus on the productive interactions between health systems and patients and providers, encompassing communication between patients and providers and among providers, and on transitions between different caregivers and settings. Third, we describe how behaviors on the part of patients and providers influence chronic disease safety in complex and reciprocal ways. Although safety threats often involve contributors within and across multiple levels from the model, to better elucidate the elements of this model we present case vignettes that highlight a particular component of the framework. These examples are drawn from actual cases that we have observed; identifying details (gender and initials) have been changed.

How Do the Community and Health System Affect Patient Safety?

Case 1. Health Information Decentralization.

Mr. J, a patient with severe degenerative joint disease who is cared for by a rural physician, is referred to an orthopedist at an urban center. He receives a chest x-ray as part of the preoperative evaluation for knee replacement. The chest x-ray shows a mass, and his knee surgery is cancelled. The orthopedic surgeon is on vacation the following month, and the radiology report is never sent to the primary care physician. Mr. J follows up three months later with his primary care physician, who learns of the chest x-ray from Mr. J. He is found to have a primary lung cancer, which is successfully removed with surgery.

Mr. J’s delayed cancer diagnosis did not cause him harm, making this event a near miss, but only because the patient himself reported the test result to his physician. In ambulatory health systems, lack of information availability is a common problem. Poor information availability can lead to delays in diagnosis and treatment, which can then lead to errors and poor health outcomes.²⁸ Studies of ambulatory medical errors identify diagnostic delays²⁸ and lack of real-time information^30,31 as important contributing factors.

Health information technology (IT), such as digitized radiology images and reports with remote-access capability, could have reduced the diagnostic delay for Mr. J. However, because ambulatory practices typically lack the personnel and resources that hospitals can provide, health IT–driven information systems remain out of reach for all but the largest ambulatory care networks.² Although strategies such as computer physician order entry and computer medication monitoring are integral to patient safety guidelines,³² they remain the exception rather than the rule. Although health IT has been shown to facilitate reporting and detection of safety problems,^33–38 both over- and underdetection of safety threats occur,^39,40 and clinical outcomes and quality of care may not improve.^37,41,42 Nevertheless, we be lieve that expanding and rigorously evaluating health IT innovations, particularly across disparate ambulatory systems, comprises a necessary step toward improving ambulatory safety.

In addition, accreditation requirements are far less stringent for ambulatory practices than for hospitals. For example, in the inpatient setting, the Joint Commission requires multidisciplinary investigation, including root cause analysis, into certain types of adverse events. Most ambulatory practices are not subject to such accreditation and do not have a mechanism for investigating adverse events, even in the unlikely event that providers submit reports. This lack of scrutiny makes understanding current safety conditions and achieving uniform safety standards a challenge. A root cause analysis of Mr. J’s near-miss situation might uncover fixable processes to deliver information. We propose careful development and implementation of safety-focused accreditation requirements for ambulatory practices.

How Do Interactions Affect Safety in Ambulatory Care?

Case 2. Transitions.

Mr. F, who has diabetes, hypertension, and heart failure, sees a primary care physician, an endocrinologist, and a cardiologist. All three adjust his medications. When he presents for a scheduled primary care visit, he does not have his medicines, so the primary care physician does not have an accurate accounting of Mr. F’s current drug regimen. Also, Mr. F did not submit to laboratory tests as requested at his prior primary care visit. His daughter, who cares for him, states that his endocrinologist had ordered laboratory tests the prior month, so she thought he did not need any more blood drawn. He reports feeling generally weak and unwell, so his primary care physician orders laboratory tests done the same day, and he is found to have dangerously low serum sodium.

Mr. F is at high risk for patient safety problems because he has multiple chronic conditions and because he makes transitions among multiple providers. Transitions among care settings and among primary care, specialty care, pharmacy, other providers, caregivers, and home care all pose a risk for adverse events.^43–45 At each point, patients must understand and carry out the plan of care, and providers must develop a plan even when lacking relevant, complete information.³⁰ The concept of a medical home with coordinated care seeks to address the risks to safety inherent in these frequent transitions.⁴⁶

For patients with multiple chronic conditions, clinicians may need to alter their medication-prescribing behavior, for several reasons. First, each additional medication increases the likelihood of patient errors in medication administration. Second, patients with multiple chronic illnesses also face increased risk for medication interactions that can lead to ADEs. Third, more general frailty may lead to more severe ADEs. For example, Mr. F’s low serum sodium may cause him to fall and sustain a serious injury, whereas a comparable abnormality may not cause as much distress in a healthier individual. Finally, because of competing effects of multiple chronic conditions on morbidity and mortality, meeting all condition-specific performance measures may lead to unnecessarily aggressive treatment goals, as shown in studies of rational screening practices.^47,48

Case 3. Inadequate Health Literacy.

Mrs. P, a patient with long-standing diabetes, has had worsening glycemic control during the past few months, although she reports being adherent to her diabetes medicines. She is referred to a diabetes educator, who calls her pharmacy and finds she has not refilled her Metformin in six months. On review, the diabetes educator finds that Mrs. P is unable to read the labels on her medication bottles because of limited literacy and misidentifies Acetaminophen, for her arthritis, for Metformin.

Mrs. P’s inability to read and correctly interpret medication labels, a common problem,⁴⁹ led to months of poor diabetes control and placed her at risk for chronic acetaminophen toxicity.⁵⁰ The adequacy of communication between patients and caregivers/providers is crucial to patient safety. Suboptimal clinician-patient communication in chronic disease care is a consequence of multiple influences at the practice and system levels, including medication labeling procedures and the communication practices of physicians and pharmacists.⁵¹

Ethnic and linguistic diversity⁵² and health literacy also influence communication and play a central role in ambulatory safety. Individuals with limited health literacy and language barriers report greater problems across a range of communication domains, including informed consent, shared decision making, and elicitation of concerns.^51,53–56 Patients with limited health literacy experience greater medication miscommunication that can pose a threat to patient safety in the context of chronic disease management.^13,51,57 Effective interventions such as visual medication schedules^14,58 and tailored medication education^59–61 should be more widely implemented for chronic disease patients.

Case 4. Patient-Physician Communication.

Mr. M. is brought into the emergency department after collapsing on the sidewalk. He is found to be hypoglycemic, and after treatment, explains that he is Muslim and fasting for Ramadan but still taking all his diabetes medicines. He did not discuss his eating patterns for Ramadan with his physician.

Although Mr. M’s religious beliefs affected his diabetes self-management, his hypoglycemic episode was preventable. His medication regimen could have been adjusted to accommodate his altered eating pattern during his religious observance. This would have required the patient to be an active communicator, conveying his religious requirements to his primary care provider in advance. In turn, this would have to be followed by appropriate medication adjustment, diet self-management on the part of the patient, and close provider follow-up.

What Patient and Provider Behaviors Enhance or Jeopardize Safety in the Ambulatory Setting?

Both patient and provider behaviors, influenced by the community, health system, and interactions in care, directly affect patient safety. As an example, ambulatory patients must perform a series of actions for appropriate medication use, including making decisions in an office encounter, obtaining a prescription, bringing the prescription to a pharmacy, receiving the medicines and instructions, taking the medication correctly at home on an ongoing basis, monitoring oneself for side effects, and following up with laboratory testing or provider visits. Problems at any of these junctures may lead to ADEs.

Case 5. Caregiving and Medication Misuse.

Mrs. S, an 80-year-old woman with hypertension, mild dementia, and arthritis, suffers from insomnia. Because she has had several recent illnesses, including a hospitalization for pneumonia, her insomnia has not been discussed at primary care visits. On a night with particularly troubling sleep problems, her son (and primary caregiver) gives her one of his own sleeping pills. In the early morning hours Mrs. S awakens confused and falls on her way to the bathroom, fracturing her hip.

Caregiving further complicates ambulatory medication management, as in the case of Mrs. S, whose son’s well-intentioned attempt to help his mother contributed to a fall and resultant morbidity from hip fracture. This case involves problems with the health system, including poor access to/knowledge of after-hours medical advice, and with communication, because a bothersome symptom—insomnia—was not discussed in recent visits. Importantly, this example illustrates how patient and caregiver errors, in this case taking someone else’s medication, can lead to harm.^7,16

Case 6. Symptom Recognition.

Mrs. K has rheumatoid arthritis, for which she is treated with methotrexate and etanercept. She presents for routine follow-up with her primary care physician, and as she is leaving, mentions that she must be approaching menopause, because she has been experiencing progressively severe headaches and “hot flashes.” Further history reveals shaking chills and frontal headache for the last month that wakes her from sleep. The patient is sent to the emergency department for further work-up, where she is found to have severe bacterial sinusitis that requires hospital admission and intravenous antibiotics.

Mrs. K did not recognize symptoms of an infection, which is a particularly important aspect of self-care in patients on chronic immunosuppressive medication such as etanercept. Although it is not possible to avoid all ADEs, there are medications that are known to cause many ADEs, including insulin^26,62 and warfarin,^14,58 and others with less common, but known, serious adverse effects, such as methotrexate⁶³ and amiodarone.⁶⁴ For these medications, symptom recognition is a crucial aspect of self-management, and appropriate communication must be the standard of care. Medication management is only one aspect of patient self-management, which also includes appropriate diet and exercise, appointment adherence, and recognition of symptoms.

Discussion

These real-life vignettes illustrate both the problems with and opportunities for patient safety in ambulatory chronic disease management. A goal of eradicating all adverse events is clearly impossible, especially with the presence of difficult-to-change factors at the community and health-system levels and the high prevalence of chronic diseases. Our aim is to highlight the many preventable events, or those in which harm to patients could have been reduced. We believe that improving patient safety could significantly reduce chronic disease morbidity and mortality.

Ambulatory Health Systems Need More Surveillance for Patient Safety Problems

Currently, we detect most patient safety problems through passive reporting at patients’ provider visits or when events lead to acute care utilization. In ambulatory settings, systems for detection of adverse events are underdeveloped and under-utilized. Unlike acute care systems, for which accreditation and regulatory agencies impose minimum standards for incident reporting and investigation, ambulatory reporting systems are of variable, often inadequate, quality.^7,26,65–68 These methods reveal the tip of the iceberg of patient safety problems; we need a clearer understanding about how patients manage their chronic illnesses at home, between visits. A recent study using between-visit telephone surveillance of diabetes patients found frequent adverse events and potential adverse events, most of which were unknown to primary care physicians but were preventable.⁹ This suggests that active surveillance in ambulatory settings would likely identify a greater number and different kinds of safety problems than do existing methods. Future studies should investigate and compare methods for detection of adverse events, including active surveillance linked to between-visit self-management support, active and passive provider-directed event reporting, and innovative use of ambulatory administrative data.

For chronic diseases, patients themselves are the key to safety. Their everyday actions to manage their health are crucial to achieving positive health outcomes and avoiding safety problems. Because of the proliferation of between-visit self-management support programs for chronic disease patients,^69,70 purposefully linking these programs with safety monitoring and promotion is a novel and promising approach. As an example, we are embarking on an innovative program to link diabetes self-management support with real-time safety surveillance to investigate adverse events prospectively. This patient-centered approach to safety can take advantage of quality improvement efforts for chronic disease care.

Providers and Policymakers Must Examine the Safety Consequences of Chronic Disease Treatment Intensification

Although failure to deliver indicated treatments can contribute to both inadequate quality of care and may represent a patient safety issue, aggressive pursuit of quality metrics may lead to inappropriate treatment intensification that actually worsens safety and outcomes.⁷¹ Evidence that adverse drug reactions are related to the overall number of medications prescribed^62,72–74 suggests that complex regimens, as well as high-risk medicines such as warfarin and insulin, should be considered a safety risk. For instance, aggressively lowering blood glucose or blood pressure, for example, in hopes of reducing risk of future complications or meeting performance goals/quality metrics, may increase adverse treatment effects such as symptomatic hypoglycemia or syncope, respectively. Yet, it is these goals that are increasingly assessed as indicators of clinician quality and as rewards for superior performance. Such quality measures may need to take into account patient capacity and characteristics such as age and comorbid illness. Moreover, any intervention to improve chronic disease care quality through aggressive treatment should incorporate rigorous safety monitoring and build in safeguards such as patient education.

All Stakeholders Must Attend to Chronic Disease Disparities to Improve Patient Safety

Finally, to have the broadest public health impact, research on safety should focus on patients with a disproportionate chronic disease burden. We need further work for those with multiple chronic diseases, with direct examination of the safety consequences of competing health demands. Because low-income, ethnically diverse, and limited health literacy patients suffer disproportionate chronic disease complications, these populations should be the focus of both research and intervention efforts.