Patient Safety in Internal Medicine

Abstract

Hospital Internal Medicine (IM) is the branch of medicine that deals with the diagnosis and non-surgical treatment of diseases, providing the comprehensive care in the office and in the hospital, managing both common and complex illnesses of adolescents, adults, and the elderly. IM is a key ward for Health National Services. In Italy, for example, about 17.3% of acute patients are discharged from the IM departments. After the epidemiological transition to chronic/degenerative diseases, patients admitted to hospital are often poly-pathological and so requiring a global approach as in IM. As such transition was not associated—with rare exceptions—to hospital re-organization of beds and workforce, IM wards are often overcrowded, burdened by off-wards patients and subjected to high turnover and discharge pressure. All these factors contribute to amplify some traditional clinical risks for patients and health operators. The aim of our review is to describe several potential errors and their prevention strategies, which should be implemented by physicians, nurses, and other healthcare professionals working in IM wards.

FormalPara Learning Objectives/Questions Covered in the Chapter

• How many are the adverse events (AEs) in Internal Medicine (IM)?

• What are the most frequent errors?

• How to prevent medication or identification errors?

• How to prevent AEs in invasive procedures in IM?

• How to prevent clinical reasoning errors?

• How to improve team working and communication among health operators in IM?

• What are the safety practices to be implemented in IM?

1 Epidemiology of Adverse Events

There are few specific studies on epidemiology of AEs in IM. Most of them are focused on particular events, such as medication, interventional procedures, or diagnostic reasoning errors.

The first historical study conducted in IM was that by Schimmel in 1960 [1]. He found that 20% of patients admitted to a university medical service in USA experienced one or more untoward “iatrogenic” episodes. Anyway, such pioneering study was not based on the current definition of AE and reported only drug reactions and untoward effects of diagnostic and therapeutic procedures — the so-called diseases of medical progresses, the price to pay for modern medical care [2, 3]. Twenty years later, Steel et al. [4] reported a rate of 36% AEs in the medical service of a teaching hospital. Then, the Harvard Medical Practice Study I [5] found a rate of AEs of 3.6 ± 0.3% (30.9 ± 4.4% of them due to negligence) in IM and 7 ± 0.5% (28 ± 3.4% of them due to negligence) in general surgery, and the Quality in Australian Healthcare Study (QAHCS) displayed an incidence of 6.6% in IM versus 13.8% in general surgery [6]. More recently, studies from the UK [7], the USA [8], Portugal [9], and Spain [10] reported an incidence ranging from 10% to 23.2%. Fatality ranges from 2% [2] to 20% [6] in the various studies. Such large variability of incidence and severity can depend on differences in AEs definition, settings (only IM wards or all medical wards), study design, and severity threshold of investigators in the adjudication of events.

Medical errors — compared to surgical ones — are more preventable (73% vs 53% [6]; 75% vs 41% [7]), and often less overt because diagnosis and time of occurrence can be less clear and multiple providers are involved [11]. They are also associated to longer hospitalizations being directly related to the time spent in hospital [1, 12]. Indeed, they have been defined “the hazards of hospitalization” [1]. They are more common and severe in the elderly [10, 12, 13], and more events can occur in the same patient [1]. Lower educational level, transfer from other institutions, associated chronic conditions, severe prognosis on admission, general functional status on admission, level of awareness on admission and at discharge, associated kidney/liver failure or impaired function on admission and at discharge, number of drugs taken (on admission, during hospital stay, and at discharge), patients’ knowledge about disease, medications, and their side effects [9] are other known risk factors for AEs in IM.

2 Most Common Errors

In IM, errors can occur in any step of inpatient journey from admission to discharge, and in any clinical process from clinical history collection to diagnostic work-up, drug therapy, invasive procedures, and so on. Further, they can occur before admission to IM and be recognized later, on the ward.

The words “error” and “AE” do not have the same meaning. AEs are “injuries caused by medical management rather than by the underlying disease or condition of the patient”. Medical errors can result or not in patient harm, but not all of them lead to AEs. Generally, only preventable AEs imply medical errors [14]. Table 17.1 displays the most frequent AEs occurring in IM, according to hospitalization phase and process.

Table 17.1 List of the most frequent AEs in internal medicine according to hospitalization phase and process

AEs in IM have been classified variously, e.g. according to the clinical process or the nature of disorder caused by AEs [7, 9] (see Tables 17.2 and 17.3). It is disappointing how the frequency of certain AEs has worsened in decades: healthcare-associated infections (HAIs) passed from 9.5% in 1960 [1] to 21.4% in 2008 [10].

Table 17.2 Types and preventability of AEs in IM [9]

Table 17.3 Types of AEs, classified according to the nature of resulting disorder [11]

When you think about the potential most frequent errors in IM, you probably think mainly about medication and diagnostic errors since medical diagnosis and therapy are its core business. Diagnostic errors —more appropriately defined as “decision-making errors”— account for 10–15% in complex disciplines, such as IM, compared to 2–5% of perceptive ones (dermatology or radiology) [15]. Medication errors are highly prevalent among older patients or patients with multiple comorbidities and polypharmacy [16], all patients typically admitted to IM. Moreover, healthcare-acquired infections are likely to be another common AE in IM, favored by intravascular catheters and immunosuppressant treatments [17].

2.1 Patient Identification Errors

Identification errors (IEs) are commonly associated with surgery, but they can occur in every setting. Many other medical errors, included in this review, such as medication or blood transfusion errors, can result from patient misidentification at the point of care as well as at registration. IEs usually affect more people. When a patient receives a medication intended for another patient, the harm is done to the patient receiving the wrong medication and to that who fails to receive the correct treatment [18]. A recent review from ECRI institute disclosed that 72% of IEs occur at the point of care and 12.6% at registration. Diagnostic and therapeutic procedures are involved in 36% and 22% of cases respectively, and consequences may be fatal [18]. Information technology amplified the problem, as IEs can generate duplicate medical records or mistaken identity. There are no specific studies on IEs in IM, but increasing staff workload and patients cognitive impairment make them a non-negligible problem.

The main barrier to IEs is cultural: the awareness of the correct identification and of misidentification consequences must be improved, so that health operators spontaneously abandon incorrect practice. Figure 17.1 summarizes what to do and not to do to prevent IEs. Technology (patient’s palm scan, bar-code wristband, radiofrequency identification system, etc.) can help but cannot substitute the role of humans. One can scan the bar-code wristband of the right patient, but administrate the drugs to another one. Patients’ education and empowerment are equally important [18].

Fig. 17.1

What healthcare operators have to do and not to do to avoid patient identification errors

2.2 Clinical Reasoning Errors

Errors in diagnostic and management process can be considered together as clinical reasoning errors (CREs) [19] or decision-making errors, as diagnostic and management reasoning can be similarly conceptualized.

According to the American National Academy of Medicine (previous Institute of Medicine), a diagnostic error is a failure to: (a) establish an accurate and timely explanation of the patient’s health problem(s) or (b) communicate that explanation to the patient. This definition includes: wrong, delayed, or omitted diagnosis [20]. The incidence of diagnostic errors varies according to definition, discipline, and research approach. For instance, 1 in 10 diagnoses are wrong (according to “secret shoppers” approach that uses “secret patients” to provide detailed, unbiased insights, and feedback on healthcare processes), 1 in 10–20 autopsies identifies major diagnostic discrepancies, 1 in 3 patients have experienced a diagnostic error (according to patients’ survey), 1 in 20 patients will experience a diagnostic error every year (according to chart review). They are the most common cause for malpractice claims [21], and about half of physicians admit at least a diagnostic error per month and perceive diagnostic errors as the most dangerous (according to physicians’ survey) [22].

The most commonly missed or delayed diseases are: pulmonary embolism and drug reaction or overdose (2.5%), lung cancer (3.9%), colorectal cancer (3.6%), acute coronary syndrome (3.1%), breast cancer (2.9%), and stroke (2.6%) [23]. Physicians overestimate their diagnostic ability: only 10% of clinicians admit they performed any error in diagnosis over the past year, but up to 40% of diagnoses about which clinicians were certain resulted wrong at autopsy [24]. Further, even when diagnosis is right, management errors can arise: 1 in 2 patients with acute or chronic diseases do not receive evidence-based therapies and 1 in 3–5 receive unnecessary and/or potentially dangerous drugs or investigations [19].

A third of CREs derive from deficits of execution (slips, lapses, or oversights in carrying out appropriate management in correctly diagnosed patients), but almost half are errors of reasoning or decision quality (failure to elicit, synthesize, decide, or act on clinical information).

Death or permanent disability result in 25% of cases, and at least three quarters of them are considered highly preventable [22].

A cornerstone of research on CREs in IM is the work by Graber et al. [25]. They analyzed 100 cases and grouped diagnostic errors in three categories:

• no-fault errors (in case of masked or unusual disease presentation or non-collaborative patient) 7%

• system-related errors (technical failure and equipment problems or organizational flaws) 19%

• cognitive errors (faulty knowledge, data gathering, or synthesis) 28%.

Coexisting system-related and cognitive errors were reported in 46% of cases. Further, wrong diagnosis was characterized by a predominance of cognitive errors (92% vs 50%), whereas delayed diagnosis by the predominance of system-related ones (89% vs 36%). Cases where discrepancy resulted from autopsy were mainly due to cognitive factors (90% vs 10%). Overall, 228 system-related factors and 320 cognitive factors, averaging 5.9 per case, were identified [26]. Among cognitive factors, faulty data gathering (14%) or synthesis (83%) resulted more frequently involved than faulty knowledge (3%) [26].

Clinical reasoning can proceed analytically or non-analytically (Table 17.4) to generate diagnostic hypotheses, investigations, and treatment. Analytical reasoning (also called “hypothetic-deductive model”) is commonly used by younger physicians or in unfamiliar or unusual cases and is based on lists of differential diagnoses and gathering of information to validate such diagnoses. Non-analytical reasoning is faster and based on mental heuristics (maxims, shortcuts, rules of thumb) or pattern recognition. In practice, physician compare current patient’s symptoms/signs with previous cases, collected through clinical experience and/or study and get the right diagnosis in few seconds [27]. One type does not exclude the other and they can be mutually used in the same patient. None of them is error-proof. If mental heuristics and pattern recognition are efficient and accurate in many situations, they can also predispose to errors, as patient’s picture does not always fit the expected pattern, because of an atypical presentation, comorbidities, or evolving diseases [28]. Another Achille’s heel of non-analytical reasoning (N-AR) are biases, constructs founded on perceptions, prejudices or ideologies, outside of critical thinking. Bias can be distinguished in internal or external to the clinicians [19] either in cognitive or affective bias [27] (see Table 17.5). Breakdowns in analytical reasoning most often derive from not following appropriate diagnostic “rules” and include: missing key data, inadequate review of existing data, deficits in medical knowledge, lacking skills in evidence-based practice and decision-making, erroneous consideration of tests value, poor supervision of N-AR [28]. At the end, also noisy environment, interruptions, high workload, fatigue, and time pressure can impair reasoning [27].

Table 17.4 Types of clinical reasoning: a comparison [27]

Table 17.5 Bias and heuristics in clinical reasoning: examples and corrective strategies [19]

Health Research & Educational Trust (HRET), Hospital Improvement Innovation Network (HIIN) team, and Society to Improve Diagnosis in Medicine (SIDM) [29] published “Diagnostic error—Change Package,” a document including a menu of strategies and concepts that any hospital should implement (improving teamwork effectiveness and diagnostic process reliability, engaging patients and caregivers, reinforcing learning system, and optimizing cognitive performances of clinicians) [29]. For this last aim, several tools are available: (a) checklists for diagnostic process such as CATCH (Comprehensive history and physical exam, Alternate explanations, Take a diagnostic timeout to be certain, Consider critical diagnoses not be missed, Help if needed) [30]; (b) mnemonic decision support tools like VITAMIN CC & D checklist (Vascular, Infection & Intoxication, Trauma & Toxins, Autoimmune, Metabolic, Idiopathic & Iatrogenic, Neoplastic, Congenital, Conversion, Degenerative); (c) lists of Red Flags; (d) electronic decision support systems like Isabel, associated with the highest accurate diagnosis retrieval rates [31]; (e) debiasing questions (Table 17.6) [32]; (f) reflective practice by the following options:

• The crystal ball experience [29]: stop and ask: “if my diagnosis was wrong, which alternatives should I consider?”

• The ROWS (Rule Out Worst case Scenario) [29]: exclude first the most severe possible diagnoses.

• The Blue and Red Team Challenge [33], borrowed from military sector, is a safe method to improve clinical decision-making in complex clinical situations. Staff is divided into two teams: the Blue Team takes clinical history, makes the synthesis and generates diagnostic hypotheses; the Red Team acts as an independent reviewer by thinking critically about the clinical picture and identifying alternative diagnoses to those presented.

• Take 2—think, do [32] is designed to improve awareness and recognition of potential errors and reduce morbidity and mortality of wrong, missed, or delayed diagnosis. Literally, it means “Take 2 minutes to deliberate diagnosis” to verify if there are situations that need a closer look or diagnosis re-evaluation (Think moment) and act (Do moment). A closer look is necessary if physician is Hungry, Angry, Late, or Tired (HALT), at risk of cognitive biases (e.g., context, framing bias) or in case of difficult patient engagement, knowledge deficit, time pressure, high-risk presentations; diagnosis re-evaluation if things are not going as planned, patient is deteriorating, response to treatment is not as expected, at shift change or discharge or in case of patient’s/caregiver’s concern. Strategies to review and challenge the diagnosis are individual strategies, i.e., Diagnostic Timeout; Team-based strategies, e.g., Red Team Blue Team Challenge; second opinion from specialist services or senior medical officer. Such approach helps to rule out the worst-case scenario, identify atypical or rare presentations, re-evaluate patients who do not improve, acknowledge patient and caregivers’ concerns, recognize high-risk patient groups, favor discussion or appropriate referral and escalation for diagnostic dilemmas, effective communication in case of care transfer.

Table 17.6 “Debiasing questions” to avoid cognitive errors in high-risk situations: what should I ask myself [33]

At the end, appropriate and effective clinical reasoning should be trained. The “twelve tips for teaching avoidance of diagnostic errors” and “ten commandments to reduce cognitive errors" can be helpful to this scope [32].

2.3 Medication Errors

Medication errors (MEs) are unintended, preventable events that can cause or lead to inappropriate medication use or patient harm [34]. You make MEs if you give the right medication to a wrong patient or the wrong medication/dose to the right patient, if you prescribe a medication to the wrong patient or without indication or when you forget to give a medication that was due. MEs are one of the most common medical errors occurring in every setting: 41.7% happen in care homes, 38.3% in primary care, and 20% in secondary care settings. It has been estimated that less than 1% cause harm to patients [35]. Associated harm is moderate in 26% of cases and severe in 2% [35]. They are also costly in terms of lives and resources [36].

MEs fall in the broadest category of adverse drug events (ADEs) that represent 5% of all AEs in high-income countries and 2.9% in low-middle income ones, according to WHO estimation [37]. ADEs are untoward, preventable or not, outcomes due to medications. If a patient has a skin rash due to an antibiotic, it is an ADE; if allergy was known, it is a preventable ADE. Preventable ADEs are formally MEs. Lastly, potential ADEs (pADEs) are MEs with the potential to cause an injury [38].

Given the well-known problem of under-reporting of ADEs, MEs affect about 4.8–5.3% of hospitalized patients with a significant variability by setting: intensive care is the most affected, whereas obstetrics the least as many drugs are prohibited [36, 39,40,41]. MEs may occur at any stage of medication process from ordering to transcription, dispensing, administering, and monitoring. About 80% happen during prescribing (39%) or nurse administration (38%), the remaining 20% during transcription and verification (12%) or pharmacy dispensing (11%) [42]. Any type of error can result from different proximal causes and a single proximal cause can lead to a variety of errors. For example, lack of drug knowledge can cause wrong choice, dose, frequency, route, or technique of administration. Wrong dose can result from lack of drug or patient knowledge, slip or memory lapses, transcription errors, and so on. Behind proximal causes there are latent causes or system failures. Leape et al. counted 16 different system failures, but the first seven have in common an impaired access to information and accounted for 78% of all MEs, whereas work and staff assignment have been associated to a broad range of errors such as slips, dose- and identity-checking, breakdown of allergy barriers [41].

Frequency of MEs/ADEs in IM has been poorly investigated. An 8-month prospective, cross-sectional study found that 89% of the patients experienced at least one ME during hospitalization, with a mean of 2.6 errors per patient or 0.2 errors per ordered medication. More than 70% of MEs happened during prescription. The most prevalent prescription MEs were inappropriate drug selection, prescription of unauthorized drugs or for untreated indications. The most involved drugs were cardiovascular agents followed by antibiotics, vitamins, minerals, and electrolytes [43].

MEs are more frequent and severe in the so-called high-risk situations due to high-risk patients and/or providers, medications, or settings. High-risk patients are younger or older, multi-morbid or chronic patients (with liver and/or renal impairment), on polypharmacy [44,45,46,47]. High-risk providers are younger or not expert providers [48, 49]. High-risk systems are hospitals delivering acute care (e.g. error rates are likely higher for drugs administered intravenously compared with other routes [50]) and high-risk medications are the so-called high-alert medications (HAMs) and look-alike, sound-alike medications (LASA). HAMs have a heightened risk of causing significant patient harm when used erroneously. They include drugs with a low therapeutic index and drugs at a high risk of harm when administered by the wrong route or at wrong dosage or when other system errors occur. The acronym A-PINCH serves as a reminder of them, it stays for Anti-infective, Potassium and other electrolytes, Insulin, Narcotics and other analgesics, Chemotherapeutic agents, Heparin and other anticoagulants. LASA are drugs with similar names or boxes [50].

Although there is no standard definition, polypharmacy is generally defined as the concurrent use of five or more medications [51], over-the-counter and complementary medicines included. It increases MEs because it reduces compliance and favors timing and/or dosing errors, duplications, or omissions. Drug–drug and drug–disease interactions, instead, increase ADEs [51]. It is particularly risky in IM as it cares for poly-pathological patients, even if internists could be more aware and cautious, as supposed by a French study [52].

Care transition is a key moment of care for several reasons, medication safety included. It occurs when a patient moves to, or returns from, home, hospital, residential care setting or simply outpatient clinics, general practitioners’ office or consultation. In care transition unintentional (changes not supported by clinical reason) and/or undocumented (motivated but not documented changes) medication discrepancies can occur [53]. They are MEs that can lead to ADEs. A mean of 1.72 unintentional discrepancies per patient have been reported at hospital admission (0.16 per patient potentially harmful) and 2.05 per patient (0.3 potentially harmful) at discharge from hospital [54].

Causes of MEs are numerous, so multiple simultaneous interventions are needed to reduce their rate and impact [36]. In recent years, information technology has been established as a cornerstone for MEs reduction. Recent meta-analysis highlighted that in hospital computerized physician’s order entry is associated with a greater than 50% decline in pADEs [55], and the use of bar-code assisted medication administration substantially reduced the rate of MEs and pADEs [56].

Medication reconciliation (MR) is recommended to avoid unintentional discrepancies between patients’ medications across transitions in care. At a minimum, medication reconciliation refers to the completion of a “Best Possible Medication History” (BPMH) and the act of correcting any unintended discrepancies between a patient’s previous medication regimen and the proposed medication orders at admission (from home or a healthcare facility, such as a nursing home), inpatient transfer (to or from other services or units, such as the intensive care unit), or discharge (to home or a healthcare facility). More advanced medication reconciliation involves inter-professional collaboration (e.g. a physician and nurse or pharmacist conducting medication reconciliation as a team), integration into discharge summaries and prescriptions, and provision of medication counseling to patients [23]. Medication reconciliation has also been bundled with other interventions to improve the quality of transitions in care, such as patient counseling about discharge care plans, coordination of follow-up appointments, and post-discharge telephone calls [24,25,26].

It refers to the completion of the BPMH and the correction of any unintended discrepancies between patient’s previous therapy and that prescribed on admission to hospital or other healthcare facility, at discharge from them or in case or transfer to other wards or settings. More advanced system of MR include inter-professional collaboration (physician, nurse, pharmacist as a team), integration of MR in discharge letters and prescriptions, medication counseling to patients. It seems that MR alone cannot reduce post-discharge hospital utilization within 30 days, but it requires to be associated with other interventions such as coordinated discharge plan, counseling about discharge plan to patients, follow-up appointments and post-discharge phone calls. Evidence shows that pharmacist involvement increase intervention’s success [57]. Beyond that there are several strategies that any operator can use to prevent MEs (Table 17.7).

Table 17.7 Individual behavioral strategies to avoid medication errors

2.3.1 Special Focus: Oxygen and Noninvasive Ventilation

Oxygen is actually a drug and, moreover, the most prescribed drug in hospitals. Oxygen is indicated in many critical conditions and is a life-saving drug, as it prevents severe hypoxemia. However, it can potentially cause serious damage or even death if it is not properly administered and managed. The National Patient Safety Agency (NPSA) published in 2009 a report of 281 incidents in which an inappropriate prescription and management of oxygen caused 9 deaths and contributed to other 35 [58]. The analysis of these events highlighted various error modes: (1) failed or incorrect prescription; (2) oxygen administration without a written prescription; (3) failure to monitor or to act in the event of altered oxygen saturation levels; (4) confusion between oxygen and compressed air or other gases, erroneous flows, inadvertent disconnection of the flow; (5) empty cylinder equipment, missing equipment. Therefore, NPSA has issued a series of recommendations to improve the safety of oxygen therapy (Table 17.8).

Table 17.8 Recommendations to improve safety in oxygen therapy and noninvasive ventilation [60,61,62,63]

Noninvasive mechanical ventilation, thanks to its potential for use outside intensive care, for example in IM, has been shown to significantly reduce mortality, the use of intubation and mechanical ventilation, especially in patients with COPD exacerbation.

A recent review [59] of AEs reported during noninvasive ventilation has shown some high-risk situations: (1) inadequate monitoring of patients unable to ask for help; (2) alarms deactivated by the staff; (3) staff not familiar with the ventilators and their proper use (e.g. if they require a CO₂ valve or not; when patients bring home appliances to the hospital); (4) implementation of a new ventilator or a new interface without training. In Table 17.8, Joint Commission International [60]/British Thoracic Society/Intensive Care Society [61] recommendations to improve the safety of noninvasive ventilation are listed.

Table 17.9 Recommendations to increase handover safety [77]

2.4 Interventional Procedure-Related Errors

The National Institute for Health and Care Excellence (NICE) defines an “interventional procedure” as a procedure used for diagnosis and/or treatment that involves [62]:

• making a cut or a hole to gain access to the inside of a patient’s body—for example, when carrying out an operation or inserting a tube into a blood vessel

• gaining access to a body cavity (such as the digestive system, lungs, womb, or bladder) without cutting into the body — for example, examining or carrying out treatment on the inside of the stomach using an instrument inserted via the mouth

• using electromagnetic radiation (which includes X-rays, lasers, gamma-rays, and ultraviolet light) — for example, using a laser to treat eye problems.

Interventional procedures most frequently carried out autonomously by the internists at bedside are: thoracentesis, paracentesis, rachicentesis, osteo-medullary biopsy, central venous accesses, joint aspirations, but literature does not provide data on their frequency. Errors during interventional procedures can cause various AEs of different severity, but apart from complication rates there substantially no data about other quality measures. For example, we know that the most common AE of thoracentesis is pneumothorax occurring in up to 39% of patients [63] (10–50% of them requiring tube thoracostomy), but we know very few about success rate, adequacy of the diagnostic specimens obtained, wait time, accuracy and completeness of clinical documentation, and patient satisfaction of thoracentesis and other procedures performed bedside on IM inpatients. On such premises, at General Hospital of Toronto an audit on procedural quality of interventional procedures was conducted in General Internal Medicine [64].

Over a 2-week period, 19 procedures (4 thoracenteses, 6 paracenteses, 8 lumbar punctures, and 1 arthrocentesis) were attempted, of which 14 at the bedside and 5 by interventional radiology. Only 7 (50%) of the bedside procedures were successful. The most common reason for failure was inability to aspirate fluid. Less than 25% of bedside procedures were done on ultrasound guidance. The majority were carried out by students and residents, but only 7 (50%) were documented as supervised. None of the operators used procedural timeouts or checklists. Over 50% of the bedside procedures were performed on evenings or weekends with less success (44% vs 60%), suggesting that procedures should be done during the daytime, when there is more availability of support and supervision. The quality of documentation was also suboptimal. Less than 50% of the procedures documented that the specific risks of the procedure were explained to the patient, how much local anesthetic was used, or what was the side (i.e., left or right). Communication with general practitioner was poor as well: only 66% of the discharge summaries included the date of the procedure and only 75% the results of the procedure [64]. Another study on lumbar puncture investigating for headache on an acute medical admission unit reported that documentation of position and cerebrospinal fluid (CSF) opening pressure was poor (42% and 32%, respectively) even if essential, and only 32% had paired serum glucose measured [65].

Procedure-related errors are due to procedural and system factors [66], such as lack of clinician comfort with performing the procedure, inadequate supplies, insufficient time, or patient factors such as body habitus or characteristics of the fluid collection such as loculation. Once more, there is good evidence that clinicians are performing fewer bedside procedures and are less confident in their bedside procedural skills [67, 68]. So, interventions able to improve safety turn out to be: ultrasound guidance, use of a procedure-specific checklist, patient identification policy and pre-procedural briefing about patient characteristics and risk factors, routine review of physician-specific procedural outcomes, periodic evaluation of operators’ competences, training through simulation, supervision until competence is consistently demonstrated and creation of dedicated teams [69,70,71], periodic assessment of procedural quality including informed consent obtained, waiting time, use of procedural timeout and sonography if needed, number of attempts, success and complication rate, diagnostic sampling quality, completeness of diagnostic tests, avoidance of waste, documentation completeness, legibility (for handwritten notes) and accuracy, wrong side errors, need for repeat procedure and patient satisfaction [64].

2.5 Communication Errors

Inter-professional communication in IM wards is complex, owing to the variety of patients’ population with changing clinical conditions and constant turnover, and multiple providers’ alternation [72]. A lot of information is exchanged every day among care providers in IM, through face-to-face (ward rounds, handover, briefing), synchronous (telephone or page), or asynchronous ways (clinical chart, text messages, emails, written handoff). Anyway, there are only few empirical studies that explore inter-professional communication in IM [73], even if effective inter-professional communication in such information-intensive environment is critical to achieve a safe and timely care.

The most common communication strategies in IM include: handover, ward rounds, clinical chart, briefing, and debriefing. In addition, there are other informal communication ways such as corridor conversation or chance hallway encounters.

2.5.1 Handoff

Up to 70% of sentinel events stem at least in part from miscommunications, often occurring during shift changes [74]. The transfer and acceptance of patient-care responsibility achieved through effective communication is technically called “handoff.” It is a real-time process of passing patient-specific information from one caregiver/team to another for the purpose of ensuring continuity and safety of care [75]. US International Joint Commission recommendations for handover are reported in Table 17.9 [75]. The most relevant is to refer to standardized handoff tools and methods (forms, templates, checklists, protocols, mnemonics, etc.). A recent review reported at least 24 different handoff mnemonics [76]. The minimum critical content to communicate to the receiver should include: (1) sender contact information; (2) illness assessment, including severity; (3) patient summary, including events leading up to illness or admission, hospital course, ongoing assessment, and plan of care; (4) to-do action list; (5) contingency plans; (6) allergy list; (7) code status; (8) medication list; (9) dated laboratory tests; (10) dated vital signs [75].

The most commonly used mnemonics are SBAR and its variants (I-SBARR, ISOBAR) and I-PASS. The former, developed in military setting to quickly pass information in command chain [77], has been adopted in healthcare with evidence for improved patient safety. Anyway, it is more suitable for emergency calls [77]. I-PASS Handoff Bundle was developed at the Boston Children Hospital and includes team training, verbal mnemonic, and structured printed tool. Medical errors fell by 40%—from 32% of admissions at baseline to 19% of admissions 3 months during the pilot study [78]. Currently, the I-PASS Mentored Implementation Program is a collaboration with the Society for Hospital Medicine funded by AHRQ, to facilitate implementation of the I-PASS Handoff Bundle in IM [79], as it is more suitable for complex patients.

2.5.2 Ward Round

According to the Royal College of Physicians (RCP) and the Royal College of Nurses (RCN), ward round (WR) is “a complex clinical process during which the clinical care of inpatients is reviewed” [80]. It is also considered “a ritual of hospital life” [81] and “the cornerstone of hospital care” [82]. Undoubtedly, it is the main moment of information exchange in IM [83], critical to ensure high-quality, safe, and timely care. However, modern hospital organization is threatening effective WR, in particular because of staff shortage. In order to “save the ward round,” RCP and RCN recently purposed to structure WR, as its standardization could warrant effectiveness and efficiency. A structured multidisciplinary WR includes four steps: (1) preparation; (2) pre-round briefing; (3) round; (4) post-round briefing. WR scheduling is not a negligible aspect to avoid overlapping with other activities (i.e. drug rounds, mealtimes, or visiting hours) or other team rounds in case of outliers. Inadequate scheduling can generate resources and efficiency issues but also safety problems, e.g. lack of the nurse responsible for the patient during WR and time wasted commuting to wards [80]. Preparation and pre-round briefing are critical to save time and resources for WR, post-round briefing to clearly delegate any task. A debrief should be conducted at the end of WR. Briefing and de-briefing are practices borrowed from military world where they are used to assign mission tasks and verify them at the end. Briefing should be well-structured, concise, focused, shared, and reported in medical chart. For bedside round, RCP and RCN purpose a structure with precise roles and responsibilities for doctors, nurses, other professionals, and patients, listing the activities that should be carried out by any of them. In this way, everyone brings his/her competencies and opinions, decisions are taken collegially, anyone is simultaneously informed, patients and/or caregiver actively participate and are timely informed about care plan [80]. That means no essential information is missed, breakdown in communication among team members and with patient or family is prevented, time and resources utilization is optimized, quality, and safety are warranted. Figure 17.2 includes a checklist for bedside round.

Fig. 17.2

Roles and responsibilities of the different health professionals during bedside round

Other subsidiary rounds are board rounds (BRs) and intentional rounds (IRs). BRs are held away from bedside, next to a white board. They should be used to facilitate patient review but cannot replace bedside round. They can be used also for post-round briefing to summarize all issues, identify and prioritize tasks, and delegate responsibilities appropriately [80]. IRs are pro-active nurse rounds to check patients at set intervals. During IRs, nurses assess patient’s experience and essential care needs (4 P: positioning, pain, personal needs, and placement). In terms of patient safety, positioning check helps to prevent pressure ulcers, personal needs (i.e., toilet) and placement of personal items checks reduce falls. Nevertheless, IRs facilitate team to organize workload [80].

2.5.3 Clinical Records

Keeping clinical records (CRs) is an integral component in good professional practice and the delivery of high-quality care. Regardless of the type of documentation (electronic or paper), a good and updated CRs allow continuity and coordination of care, aid informed decision-making, avoid repetition of tests or other investigations, improve communication between the various health professionals and improve time management. Bad CRS misinform healthcare professionals and patients, prolong hospitalization, jeopardize patient care leads to serious incidents and increase medical-legal risk [84]. Figure 17.3 summarizes what to do and not to do to keep good medical records.

Fig. 17.3

What healthcare operators have to do and not to do to keep good clinical records

3 Safety Practices and Implementation Strategy

According to the Agency for Healthcare Research and Quality and the National Quality Forum “a Patient Safety Practice is a type of process or structure whose application reduces the probability of adverse events resulting from exposure to the healthcare system across a range of diseases and procedures” [85].

In 2001 [86] and 2013 [85], an international panel conducted an evidence-based assessment of patient safety strategies (PSSs). The PSSs were categorized according to the following aspects: frequency and severity of the problem addressed, strength of evidence of the effectiveness of the safety strategy, the evidence or potential harmful consequence of the safety strategy, an estimation of implementation difficulties and costs. It categorizes each PSS according to the following: the scope of the underlying problem that the PSS addresses (its frequency and severity); the strength of evidence about the effectiveness of the safety strategy; the evidence or potential for harmful consequences of the strategy; a rough estimate of the cost of implementing the strategy (low, medium, or high); and an assessment of the difficulty of implementing the strategy. As a result of this process, 10 PSSs were identified as “strongly encouraged” and other 12 as “encouraged” for adoption [85].

Here, we report some safety practices relevant to IM, most of them included in the list of strongly encouraged or encouraged for adoption [87].

3.1 Prevention of Age and Frailty-Related Adverse Events

Falls. The rate of falls in acute care hospitals varies from 1 to 9 per 1000 bed-days. The first effective strategy relies on the timely recognition of patients with risk factors for falls (Table 17.10) [88]. The National Institute for Health and Care Excellence (NICE) recommends to regard as the population at risk all inpatients older than 65 and those between 50 and 64 who are identified as being at high risk of falling [89]. Actually, some tools are available to discriminate between high-and low-risk patients, but they may show limitations in specific populations. Morse Falls Score (MFS) and STRATIFY Score are the two most widely validated tools. However, they were not judged to be diffusely adopted and generate greater benefits than nursing staff clinical judgment [90]. NICE guidelines do not recommend any predictive score [89]. Besides, various assessments and interventions should take place (Table 17.11): (1) all aspects of the inpatient environment —including flooring, lighting, and furniture— must be identified and addressed; (2) high-risk patients should be considered for multi-factorial evaluation in order to timely identify cognitive impairment, incontinence, fall history, medications (Table 17.12) or health problems increasing the risk of falls, unsuitable footwear, and visual impairment. There is a high-quality evidence that multicomponent interventions can reduce risk for in-hospital falls by as much as 30% [91]. The optimal bundle is not clearly defined but relevant components are: patients risk assessment, patient and staff education, bedside signs and wristband alerts, footwear advice, scheduled and supervised toileting, and medication review [91]. In particular, patients’ education should include exhaustive oral and written information to patients/caregivers —taking into consideration the patient’s ability to understand and retain this information— about (1) patient’s risk factors for falls; (2) how to call the nurse as well as when to ask for help before moving from or around the bed; (3) when and how to raise bed rails; (4) other interventions aimed at addressing individual risk factors.

Table 17.10 Risk factors for falls in hospitalized patients [90]

Table 17.11 External and internal factors associated with falls [182]

Table 17.12 Drugs increasing the risk of falls [182]

Harms due to interventions have not been studied systematically, but they may include an increased use of restraints and sedatives and decreased patients’ mobilization [91].

Key factors for a successful implementation of such multicomponent interventions include: leadership support, engagement of frontline in the design of the intervention, multidisciplinary committee, pilot-testing the intervention, and changing nihilistic opinions about falls [91].

Wandering. It refers to two different, sometimes associated, behaviors: (1) the tendency of nursing home residents or hospital inpatients to persistently walk, spatial disorientation, or a combination of both [92]; (2) a situation in which a subject with dementia has become lost in the community. Although not all subjects with cognitive impairment exhibit wandering behavior, all are at risk for wandering away from the care setting and becoming lost [93].

The first measure to prevent wandering consists of an accurate assessment of patient’s diseases impairing cognition such as Alzheimer’s disease, fronto-temporal dementia, Lewy body disease, multi-infarct dementia, and delirium, on admission. In such cases, supervision is pivotal to reduce wandering-related problems [94] and should allow an immediate identification of patients at risk (e.g. through colored wristbands, armbands, or gowns), strategies providing an intensive surveillance (i.e. rooms close to the nursing station so that can be easily controlled by nurses and patients cannot go out without passing through it), and engagement of family members. This latter can play an important role during hospitalization as a familiar voice or face can decrease fear and agitation of the patients, thus reducing the patient’s willing of wandering. Other strategies may include the avoidance of rooms near elevators, stairs, or exit doors as patients with cognitive impairment tend to respond to what they see around them. Placing clothes, shoes, and suitcases out of the patient’s view can help as well. Finally, electronic monitoring could represent a big help, installed in the division of a hospital or a nursing home and potentially linked to local law enforcement agency, such as in the Project Lifesaver technology (https://projectlifesaver.org/).

On the other end, inappropriate building organization, overworked and under-resourced system, and limited staff knowledge of these problems may represent risk factors for patients’ wandering [95, 96].

Bed entrapment occurs when a patient is being caught, trapped, or entangled in the bed rails, mattress, or bed frame of a hospital bed [97]. Many health conditions can favor this event, such as cognitive and communication impairments, frailty, agitation, uncontrolled pain, uncontrolled body movements, and bladder and/or bowel dysfunction. Healthcare professionals should perform a patient’s evaluation to identify those at risk and monitor them by concentrating on the following elements: mental status, disease-related reasons for a reduced mobility capacity (obesity, neuromotor deficits), prior long bedridden period, risk of fall and fall-related injuries, urine/fecal incontinence, and the paradox effect of certain drugs.

In order to prevent this event, it is very important for all medical staff to familiarize with the areas of the bed where patients are most often entrapped (Fig. 17.4 and Table 17.13) [97]. These areas account for 80% of entrapment accidents occurring in the hospital. The US Food and Drug Administration (FDA) provided some precise indications for the sizes of the different parts of the bed aimed at reducing as much as possible these accidents. For instance, in order to avoid trunk, head, and neck to be blocked in the bottom part of the bed, mattresses should cover completely this area and resist to patient’s movements and weight. Similarly, entrapment risks in the empty spaces between rails should be avoided. In Table 17.13, requirements for the size of the different bed areas are provided [98].

Fig. 17.4

Areas where patients are most often entrapped. Zone 1: between the headboard or footboard and the mattress; zone 2: under the rails; zone 3: between the rail and the mattress; zone 4: under the ends of the rail; zone 5: between the 2 bed rails; zone 6: between the end of the rail and the edge of the headboard or footboard; zone 7: within the rails [100]

Table 17.13 Areas of the bed at risk for entrapment and recommendations from the US Food and Drug Administration (FDA) [99]

Aspiration pneumonia is considered as a continuum including community- and hospital-acquired pneumonias. However, data of in-hospital aspiration pneumonias are lacking as solid diagnostic criteria are not available [99, 100].

An important step to face this dangerous complication is represented by the recognition of risk factors (Table 17.14). Indeed, patients presenting with many risk factors have a 9- to 13-fold increased risk of death and adverse outcomes [101]. Compared to patients with community-acquired pneumonia, those at risk for aspiration experienced a 70% increased risk for 1-year mortality, a 3-fold risk for recurrent pneumonia, and a 1.5-fold risk for re-hospitalization [101].

Table 17.14 Risk factors for aspiration pneumonia [103,104,105]

Since most of the elderly patients admitted to IM are assuming a long list of drugs, a great effort should be done to avoid sedatives, hypnotics, antipsychotic agents, and anti-histamines, if possible [102]. Additionally, patients with dysphagia, especially those affected by a previous stroke or a neurodegenerative disease, can benefit from speech and swallowing evaluation, before allowing feeding [103]. Oral feeding should always be preferred to enteral tube feeding using a mechanical soft diet with thickened liquids, avoiding pureed food and thin liquids. However, when enteral feeding is unavoidable, patients should be positioned in a semi-recumbent and anti-Trendelenburg position to reduce the chance of gastric aspiration/regurgitation. In patients with dysphagia, it is helpful to consider a nutritional rehabilitation, during which swallowing exercises and early mobilization may reduce risks of aspiration and/or recurrences [104, 105]. While the effectiveness of the nasogastric tube and the post-pyloric feeding is controversial, the use of angiotensin-converting enzyme inhibitors (as anti-hypertensive drug) and cilostazol (as an anti-platelet drug) acting on substance P and bradykinin and improving cough and swallowing reflexes showed more consisting results [106,107,108].

Oral hygiene may represent an important preventive action in non-ventilated patients: it has been demonstrated that chlorhexidine or mechanical oral cleaning reduce up to 60% risk of aspiration pneumonia [109]. However, it is important to remember that chlorhexidine can be toxic if aspirated into the lungs, especially by ventilated patients. The association of oral care to supplemental nutrition also demonstrated to lower aspiration pneumonia [110]. Anyway, a comprehensive oral care program (manual tooth, gum brushing, chlorhexidine mouthwashes, and upright positioning during feeding) evaluated in a cluster-randomized controlled trial conducted among nursing home residents showed a higher number of pneumonias/lower respiratory tract infections in the intervention group [111]. On the other hand, a short course (≤24 h) of prophylactic β-lactam antibiotics was shown to reduce the risk of aspiration around the time of endotracheal intubation [112].

Delirium is a neuropsychiatric syndrome characterized by altered consciousness and attention with cognitive, emotional, and behavioral symptoms. It occurs among hospitalized patients—mainly in elderly frail people—at a rate from 14% to 56% and increases morbidity and mortality [113]. In this condition, multiple risk factors have been identified so that suggested intervention is obviously multicomponent. Evidence shows that they are effective in preventing delirium onset in at-risk patients in a hospital setting, without significant associated harms but it is insufficient to identify which multicomponent interventions are the most beneficial, and which components within a program provide the most benefit [114, 115]. The aim of primary prevention is to prevent physiological derangements by early mobilization, good hydration, sleep enhancement, family and caregiver involvement, in addition to physiotherapy and rehabilitation, as summarized in Table 17.15.

Table 17.15 Multicomponent non-pharmacologic approaches to prevent delirium (adapted from [116])

Since it is usually triggered by different factors, prevention strategies need to be reassessed during hospital stay [114].

Approaches including the education of nursing aides and caregivers, music therapy and psychotherapy gave no definitive results [114].

The main recently published studies on pharmacological approach are summarized in a review by Oh et al. [114]. In general, antipsychotic drugs did not demonstrate any clear benefit in preventing delirium [116], similarly to cholinesterase inhibitors, ketamine, melatonin, and melatonin-receptor agonist (ramelteon) [117, 118]. Hence, there is a lack of support in using drugs for prevention or treatment of delirium, especially when considered as a unique entity.

At the end, if non-pharmacological strategies were proved to be effective on delirium onset, no convincing impact was provided for hospital mortality, 6-month mortality, or institutionalization. As well, frailty, as a key predictor of outcomes, was not taken into consideration [119].

3.2 Prevention of Healthcare-Associated Infections

Healthcare-associated infections (HAIs) represent a relevant problem for hospitalized patients all over the world. Some 3.2 million patients in Europe suffer every year from HAIs, of which nearly one third is considered preventable [120].

Many preventive strategies may help in reducing the spreading of HAIs [121]. For instance, patients coming from the intensive care unit to IM should be screened if they present with neutropenia, diarrhea, skin rashes, known communicable disease, or if they are known carriers of an epidemic bacterial strain. The recognition of risk factors, listed in Table 17.16 may help in reducing HAIs, too.

Table 17.16 Common risk factors increasing the risk of HAIs [122, 123]

As hands are the most common vehicle for transmission of infections, hand hygiene is the single most effective measure to prevent the horizontal transmission of infections among hospitalized patients and healthcare personnel. In 2003, World Health Organization promoted a world challenge on this topic, introducing the five moments for hand hygiene, two before and three after approaching the patient: (1) before touching the patient in order to protect him/her from germs carried on healthcare personnel’s hands; (2) before aseptic procedures to protect the patient against germs, including the patient’s own ones; (3) after body fluid exposure; (4) after touching the patient; and (5) after touching the patient’s surrounding (these three latter moments are intended to protect the personnel and the environment from the patient’s germs) and two methods, with water and soap or alcohol-based solutions [122].

In addition, standard precautions include preventive measures that should always be used, irrespective of a patient’s infection status. Sterile gloves should be worn after hand hygiene in case of sterile procedures or exposition to body fluids. It is important not to wear the same gloves when caring for more patients, remove them and wash hands after caring for a single patient. Wearing gown, mask, and eye protection/face shield is very important to avoid soiling clothing and skin during procedures potentially delivering body fluids [122].

In patients known or suspected to have airborne, contact or droplet infections (M. tuberculosis, H. influenzae, varicella zoster virus, herpes virus among others), additional precautions should be followed.

For airborne infections, isolation with negative-pressure ventilation is preferable. Additionally, all people entering the room, including visitors, must wear respiratory protections (such as the disposable N-95 respirator mask).

For contact infections, single use patient-care equipment is recommended. If unavoidable, adequate cleaning and disinfection before using to another patient is mandatory. As well, the movements of the patients across different wards should be limited.

In droplet infections, the patient should be isolated and his/her movements limited, while respiratory protections must be worn when entering the isolation room. Additional specific strategies to prevent specific nosocomial infections have been reported by Mehta et al. [123].

Finally, environmental factors cannot be neglected. Adequate cleaning and disinfection are important, especially when considering the patient’s closest surfaces, such as bedrails, bedside tables, doorknobs, and equipment. The frequency of cleaning should be as follows: surface cleaning twice weekly, floor cleaning 2–3 times/day, and terminal cleaning after discharge or death. Central air-conditioning systems should ensure that air recirculates through appropriate filters (air should be filtered to 99% efficiency down to 5 μm). Isolation facility should include both negative- and positive-pressure ventilations. Alcohol gel dispensers should be positioned at the entry of every rooms and near entrance/exit for health operators, patients, and visitors.

3.3 Prevention of Venous Thromboembolism

The hospitalization for an acute condition is responsible for an eight-fold increase in the thrombotic risk and accounts for nearly 25% of all thromboembolic events [124]. However, risk stratification of patients admitted to IM is often complicated by their high heterogeneity [125, 126]. For this purpose, the Padua Prediction Score has been implemented and validated by Prandoni et al. [126]. It includes 11 thrombotic risk factors and identifies patients at high or low risk for venous thromboembolism (VTE) (Table 17.17). Patients with a score <4 (nearly 60% of the patients) are at low risk, while those with a risk score ≥4 (nearly 40%) have a high risk. Indeed, in the 3-month follow-up period, the incidence of VTE without any prophylaxis in the low-risk group was 0.3%, while the incidence in the high-risk group was 11% (hazard ratio HR 32.0, 95% confidence interval 4.1–251.0). Based on these findings, the Padua Prediction Score was recommended as a tool for the identification of high-risk patients requiring thromboprophylaxis [125]. Anyway, the hemorrhagic risk should also be considered. In the study by Prandoni et al., major or clinically relevant bleeding complications were found in 1.6% of high-risk patients receiving pharmacological prophylaxis although all bleeding complications were non-fatal [126]. In another study, active gastroduodenal ulcer, prior bleeding within 3 months, and low platelet count (<50,000/mm³) were recognized as the strongest independent risk factors for bleeding [127]. Other bleeding risk factors included age >85 years, male sex, hepatic or renal failure, intensive care unit stay, central venous catheter, rheumatic disease, and cancer. All these factors have been integrated in a score for bleeding risk stratification (IMPROVE score), highlighting that more than a half of the major bleeding events were experienced by patients with a score ≥7 [127].

Table 17.17 The Padua Prediction Score [127]

Combining thrombotic and hemorrhagic risk assessments, pharmacological and non-pharmacological measures can be adopted to safely reduce in-hospital VTE [128].

Current evidence is concordant in recognizing a similar efficacy of low-molecular-weight heparin (LMWH) and low-dose unfractionated heparin (LDUH) in patients hospitalized in the medical setting although LMWH is more likely to be associated with a lower risk of bleeding. Fondaparinux, the only selective inhibitor of factor Xa approved for the treatment and prevention of thrombosis, showed a similar performance compared to heparin both in terms of thromboprophylaxis and risk of bleeding [125]. For patients with an increased risk of bleeding, alternative treatments, such as graduated compression stockings, intermittent pneumatic compression, and venous foot pumps, all aiming at reducing venous stasis by inducing the movement of blood from superficial to deep veins through the perforator veins are recommended [125].

Since IM usually receives a great number of patients often showing particular features (elderly, obese or underweight people, impaired kidney function, cancer), these specific populations need different managements [129].

Elderly patients present differences in terms of pharmacokinetics and an increased risk of bleeding, compared to the general population [130]. Further, older patients (>80 years) show a ten-fold risk increased risk for VTE compared to younger ones. Indeed, in the MEDENOX study, enoxaparin was greatly effective in reducing the risk of VTE in patients >80 years hospitalized in medical wards [131].

Obesity and overweight are recognized risk factors for VTE. The main concern is to modify or not the dosages to get the same efficacy in such conditions. A study conducted in a medical ward in the USA tested the 0.5 mg/kg/day enoxaparin dosage in obese patients showing its feasibility and efficacy and, at the same time, the absence of any bleeding event, symptomatic VTE, or dangerous thrombocytopenia [132]. Some differences arose in a study among patients undergoing bariatric surgery [133] underlining potential differences in terms of absorption among the different formulations of LMWH. For this reason, for obese patients, dosages may need to be modified according to the drug used.

In patients with kidney disease, LMWH and fondaparinux clearance is reduced and a modification of the dosage is required. Usually, LMWH can be used at the dosage indicated for thromboprophylaxis with a limited risk of bioaccumulation in patients with kidney disease treated for a limited period of time [134]. LDUH can be a valid alternative in patients with advanced kidney disease. Prophylactic doses of fondaparinux must be reduced when kidney function is severely impaired: 1.5 mg/day when estimated glomerular filtration rate (eGFR) is 20–50 mL/min/1.73 m². Fondaparinux is not recommended when eGFR is below 20 mL/min/1.73 m² [135].

Patients with active cancer are known to be at increased risk of arterial embolism and VTE as well as bleeding events. Although treated for a long time with LMWH, recently direct oral anticoagulants have been found to be effective in reducing the risk of VTE and arterial embolism in many large randomized clinical trials. With this regard, an exhaustive report on these therapeutic strategies can be found in a recent review by Mosarla et al. [136]. Direct oral anticoagulants, however, are not yet approved for the prophylaxis of venous thromboembolism in these patients, but only in secondary prevention.

3.4 Prevention of Pressure Ulcers

Complications from hospital-acquired pressure ulcers cause about 60,000 deaths and relevant morbidity and resources consumption every year in the USA. Diabetes, obesity, and older age are known risk factors [137].

Moderate-strength evidence suggests that implementing multicomponent initiatives for pressure ulcer prevention in acute and long-term care settings can improve processes of care and reduce pressure ulcer rates [137].

Interventions usually address impaired mobility and/or nutrition and/or skin health. Using support surfaces, regularly repositioning the patient, optimizing nutritional status, and moisturizing sacral skin help to prevent pressure ulcers, along with initial and periodic risk stratification and personalized care for high-risk individuals. Many different pressure ulcer risk assessment tools are used in clinical practice (i.e. Braden, Norton, Exton-Smith, Waterlow, Knoll, …), but a recent Cochrane review was unable to suggest that the use of one tool over the others because of low or very low certainty of available evidence [138]. Multicomponent interventions typically include 3–5 evidence-based practices that “when performed collectively and reliably, have been proven to improve patient outcomes” [139]. Further, experts recommend to pay attention to organizational and care coordination components [140, 141]. Organizational components include selecting lead team membership, establishing policies and procedures, evaluating quality processes, educating staff, using skin champions, and communicating written care plans. Care coordination components include creating a culture of change and establishing regular meetings to facilitate communication, collegiality, and learning [137].

Key components of successful implementation efforts include: simplification and standardization of pressure ulcer-specific interventions and documentation, involvement of multidisciplinary teams and leadership, designated skin champions, ongoing staff education, and sustained audit and feedback [137].

3.5 Clinical Monitoring by Early Warning Scores

Many hospitalized patients experience vital signs deterioration before cardiac arrest, unanticipated intensive care unit admission or unexpected death [142, 143]. Indeed, one or more aberrant vital signs can be detected by nurses or physicians in 60% of cases before the adverse event [144]. A rapid recognition of these antecedents and an appropriate treatment can prevent further deterioration so avoiding the development of the adverse outcomes. Several studies suggest that the triad of (1) early detection, (2) timeliness of response, and (3) competency of the response is crucial for patient’s outcomes [145,146,147]. According to these considerations, the use of the so-called early warning scores (EWS) has been widely implemented by hospitals to efficiently identify and treat patients who present with or develop acute illness [147, 148]. Although different and heterogeneous EWS exist, they are characterized by few key features. First, they require a systematic method to measure simple vital signs at the right intervals in all patients to recognize those with clinical deterioration. The assessment of vital signs need to be simple and usable by all healthcare professionals after an appropriate training. Second, clear definitions of the urgency and of the appropriate clinical response are necessary. The trigger for the clinical response should not be too sensitive in order to avoid alerts but it also should not be so insensitive that it never leads to system response activation [149]. In the EWS, the points for the final score are allocated for each physiologic parameter according to how much it deviates from a predefined normal range, so that a higher score corresponds to greater patient’s deterioration. So, clinical response can be adapted in terms of urgency and provider’s level of expertise, ranging from the increase of vital signs monitoring to the activation of rapid response team. The vital signs considered in each EWS typically include pulse rate, breathing rate, blood pressure, level of consciousness and temperature [150]. There is, however, variability in other parameters included (e.g. pain, level of respiratory support, urine, age), in weights assigned, and in thresholds for triggering the response. In Table 17.18, the chart of National Early Warning Score (NEWS) used in the UK is reported, as an example [151]. Another important issue to consider is the frequency of vital signs monitoring. Ideally, it should be done frequently enough to identify patient’s deterioration at a time that allows interventions to improve outcomes. There is no evidence that continuous surveillance has a positive effect on mortality [152, 153]. Moreover, although an increase in monitoring frequency leads to a higher detection of events, it is also associated with a rise in expense and workload [149]. Thus, it is necessary to find a balance between patient’s safety and available resources. According to evidence, patients at low risk should be monitored at least twice daily, whereas an increase in assessment frequency is required when EWS raise [154]. The appropriate responses to EWS can be described with an escalation protocol, in which at every threshold corresponds an action (see Table 17.19). Providers at every level of the chain have to operate according to their competences and skills. They have also to call medical emergency team (MET) when it is indicated by the protocol. Several studies, however, reported omission to call MET in 25–42% of cases in which patients presented calling criteria [155, 156]. Reasons for non-adherence to protocol include negative attitude toward MET, staffs’ confidence in their own ability, fear to appear incompetent or of criticism by the MET [155,156,157,158]. Ongoing education and training in the use of EWS is essential for all healthcare staff involved in the assessment and monitoring of acutely ill patients. A standardized system, jointly to a diffuse knowledge of it, is essential to achieve the aim of a rapid recognition of patient’s deterioration, an appropriate clinical response and a favorable outcome.

Table 17.18 National Early Warning Score (NEWS), adapted from [153]

Table 17.19 Clinical response to NEWS trigger, adapted from [153]

3.6 Sepsis Bundles

The mortality rate for severe sepsis and septic shock remains a major concern in clinical practice [159]. The Surviving Sepsis Campaign (SSC) is a joint collaboration of the Society of Critical Care Medicine and the European Society of Intensive Care Medicine created in 2002 to increase sepsis awareness, improve early diagnosis, increase the use of appropriate timely care, develop guidelines and spread them, in order to reduce morbidity and mortality for sepsis. Sepsis bundles were presented for the first time in the SSC Guideline for the management of severe sepsis and septic shock in 2004 [160]. They were created to bring guidelines key elements to clinicians’ daily practice [161]. Indeed, a bundle is a small and straightforward set of evidence-based practices that, when performed altogether, have been proven to improve outcomes [162]. Hospitals that have successfully implemented sepsis bundles have consistently shown improved outcomes and reductions in healthcare spending [163]. Over the years, sepsis bundles have been revised according to most recent scientific evidence [164, 165]. The most recent version is the hour-1 bundle, published in June 2018 [166]. Sepsis is a medical emergency. Early recognition and prompt management in the first hours after its development improve the survival [167]. Accordingly, the aim of hour-1 bundle is to begin sepsis management and resuscitation immediately although some of the actions require more than 1 h to be completed.

The hour-1 bundle includes five key steps:

• 1. Measure lactate levels and re-measure if initial lactate is >2 mmol/L. Lactate is a surrogate for tissue perfusion measurement [168]. Lactate-guide resuscitation has been shown to reduce mortality in randomized control trials [169, 170]. So that, if initial lactate is elevated (>2 mmol/L), the measure should be repeated within 2–4 h and the treatment should be based on its values with the aim of normalizing lactate.

• 2. Obtain blood cultures prior to antibiotics administration (at least two sets, aerobic and anaerobic). If obtaining blood cultures is difficult, however, do not delay antibiotic treatment beginning. The identification of pathogens improve outcomes, but can be difficult to obtain after antimicrobial treatment for the rapid sterilization of cultures [171].

• 3. Administer broad-spectrum antibiotics. The antimicrobial treatment should be started empirically with one or more intravenous broad-spectrum antibiotics. Therapy should be narrowed once pathogen is identified.

• 4. Begin rapid administration of 30 mL/kg of crystalloid fluids in case of hypotension or lactate ≥4 mmol/L. Fluid resuscitation should be started immediately after the recognition of sepsis signs. The use of colloids did not show any clear benefit and it is, therefore, not recommended by guidelines.

• 5. Administer vasopressor for hypotension during or after fluid resuscitation, in order to achieve a mean arterial pressure ≥65 mmHg.

All these actions must be initiated within 1 h from “Time Zero,” defined as the time of triage in the Emergency Department or, in case of sepsis presenting in another care location, from the earliest chart annotation consistent with elements of sepsis or septic shock.

A successful treatment of sepsis and septic shock require the collaboration of all healthcare professionals. The role of nurses is particularly important because they interact constantly with patients and they can provide early recognition of sepsis and implement a rapid clinical response [172]. Education programs on sepsis screening and hour-1 bundle should be strongly recommended for the entire medical staff. The website survivingsepsis.org provides resources and tools to improve sepsis knowledge.

3.7 Safe Management of Outlier Patients

“Outlier” or “out-lying hospital in-patient” is a patient who, is admitted wherever an unoccupied bed is, because of unavailability of hospital beds in his/her clinically appropriate ward [173, 174]. In such case, clinical management is on charge of physicians of the clinically appropriate ward (generally IM ward), but care is delivered by nursing staff of the hosting ward (often a surgical ward). Outliers phenomenon involve commonly medical patients in countries with a public health system that faced hospital beds cuts, over the last decades. Outliers represent about 7–8% of all admissions every year [173]. They are the other neglected face of hospital overcrowding. From a patient safety point of view, they have been defined, according to Reason’s Swiss cheese model, “a latent condition which may underpin adverse events.” Identification errors, missed or delayed diagnosis and treatment, HAIs, delirium and falls could be amplified by outlier status, due to delay between admission and medical evaluation, discontinuity of care, errors or delay in tests request/execution, inadequate communication between ward-teams, less familiarity with monitoring and treatment by hosting team [174]. Despite their compelling nature, they have been poorly studied. Available evidence shows a trend to increase in-hospital mortality and hospital readmission, but presents many serious limitations [174]. Also evidence-based guidelines to safely manage outliers clinical risks are still lacking. Only some bed management policies, formulated mainly by NHS Trusts across the UK [175, 176] contain some indications to ensure safety, dignity, and duty of care for both patients and staff involved in the care of outliers. As an example, that from Portsmouth Hospitals NHS trust recognizes that the best choice is not to admit to off-service units, but when unavoidable, the risk for patients and staff need to be minimized. It recommends not to admit to off-service units directly from emergency department or acute medicine, except in rare cases. It prescribes to rate patients’ suitability to be moved to other units, with a score (RAG) based on clinical and mental health needs, level of acuity and dependency and clinical capability of the receiving area. RAG must be assessed within 24 h from admission and reviewed every day. Further, outliers must be placed in the same level of care and treatment that they would receive if cared in their appropriate unit. They must be reviewed by medical and/or nursing teams from their clinically appropriate unit daily. Patient treatment plans must be updated including pending investigations and discharge plans carefully documented in the patient’s health records. The number of bed moves during each patient’s stay must be minimized. Relatives must be informed of every movement and patients must be involved in decision by signing an informed consent [177].

4 Case Studies

4.1 Case Study 1

Female, 36 y-o, immigrant, unemployed, living with her husband and a 6 y-o daughter. Access to Emergency Room (ER) at 5.30 p.m. for left flank pain and hematuria. Previous history of kidney stones. Giordano’s test positive. Her general practitioner suggests hospital admission for alcohol withdrawal. Blood tests reveal increased neutrophils, c-reactive protein and transaminases; abdominal US scan shows left hydro-nephrosis but not signs of liver damage. After 5 h, she is discharged with a diagnosis of hypertransaminasemia in chronic alcohol abuse. Left renal colic. ER physician says she preferred go back home to fix her daughter tonight and will come back tomorrow. Twelve hours later, she is back to ER. ER physicians writes: “the patient comes back for left flank pain”. Her general practitioner contacted social and psychiatric services. She remains in the ER until 5.00 p.m. without clinical nor laboratorial re-evaluation. Then she is admitted to a medical ward for bilateral renal colic and alcohol abuse. At 9 p.m. onset of worsening psychomotor agitation, treated by diazepam, gabapentin, vitamin B6, and fluids. At 8 a.m., she receives the first dose of antibiotics (i.v. piperacillin/tazobactam). At 9 a.m., nurse reports hypotension (90/60 mmHg) and low peripheral oxygen saturation (92% room air); instead physician writes in medical record “inappropriate admission,” withdrawal syndrome in chronic alcohol abuse. At 2 p.m., morning shift physician hands off the patient saying she is going home because she rejects treatment. During the afternoon, psychomotor agitation worsens so that the treatment with fluids and oxygen is compromised and relatives are asked to provide assistance to her. She receives multiple administration of i.v. midazolam. At 8 p.m., she has cardiorespiratory arrest. She is resuscitated and transferred to intensive care unit. A diagnosis of post-anoxic coma and septic shock by Escherichia coli is made and the patient dies after 20 days without ever regaining consciousness.

4.2 Case Study 2

A 78-year-old man, previous gastric ulcer and depression, affected by metastatic colon cancer in home palliative care, was admitted to IM ward on December 27th at 1.00 a.m., after rejecting hospice admission to die at home, just the day before. He was on transdermal and sublingual (breakthrough cancer pain) opioids, intravenous opioids, haloperidol, and hyoscine (elastomeric pump). He died about 20 h later. Ten days after, his wife and son made a claim for bad assistance. They complained that their relative was removed from sedation, so he was awake in the grip of its devastating pains; his pain was not asked or evaluated; no painkillers were given. They were told by the nurses: “We can’t do more than that. Sedation is a matter of anesthesia.” On the contrary, electronic medical record reported that patient was unresponsive to any stimulus since admission; sedation was not interrupted; intravenous opioids dose was progressively increased; pain evaluation was frequent and pain control was achieved in few time. Health operators declared also that his relatives were allowed to stay with him until the end and any their desire such as music listening was satisfied. Why so different perceptions?

Despite of technical expertise and some human compassion, audit disclosed communication failure, and inappropriate setting (acute care ward). First of all, ward team missed medication and care plan recognition with palliative doctors, and, most of all, it did not effectively take care of family concerns and expectations. Health operators did not explore family feelings, did not provide frequent and punctual information about what was done and reassurance about their beloved clinical condition, in particular unconsciousness.

4.3 Epicrisis and Recommendations

4.3.1 Clinical Case 1

• 1. Be aware of Medical mimics or secondary psychoses, medical conditions mimicking psychiatric disorders, especially in patients with previous psychiatric history.

• 2. Remember that infections, trauma, autoimmune, metabolic, neurological diseases, and pharmacological withdrawal can present with psychiatric symptoms, from psychomotor agitation to anxiety, depression, dementia, or apathy.

• 3. Think about medical mimics in case of: patient over 40 years and no previous psychiatric history, no history of similar symptoms or worsening of previous symptoms, family concern, chronic comorbidities, history of head injury, change in headache pattern, worsening after antipsychotics or anxiolytics, history of changing psychiatric diagnoses over time, difficult or unlikable patient, polypharmacy, abnormal autonomic signs, visual disturbance, visual, olfactory or tactile hallucinations, nystagmus, illusions, speech deficit, abnormal body movement [178].

• 4. Have a complete medical and psychiatric history, an exhaustive review of systems to identify symptoms/signs suggestive of medical diseases, review of any drug prescription, over-the-counter and alternative medications included, a careful mental status examination, diagnostic tests for diseases known to mimic psychiatric disorders (look for head trauma, syphilis or hypothyroidism, glucose or electrolyte or blood gases alterations, sepsis, etc.)

• 5. Avoid incorrect assumptions (patient triaged as psychiatric, is psychiatric; patient with psychiatric history, has only psychiatric disease; young patients suffer from functional disorders; abnormal vital signs are due to mental/emotional state) and pitfalls (cursory history from limited sources, incomplete review of system, incomplete physical and neuropsychiatric exam, failure to review medications) [179].

4.3.2 Clinical Case 2

• In end-of-life care, ensure skillful communication with patients and families.

• Define and share with patient and/or family realistic goals of care.

• Pay attention to understanding the patient’s and family’s concerns besides competent symptom management [180].