Patient Safety and Quality Improvement

Summary

• The WHO estimates that even in advanced hospital settings one in ten patients receiving health care will suffer preventable harm [1].
• Surgery carries a disproportionate share of it: in resource-rich countries, problems associated with surgical safety account for half of the avoidable adverse events that result in death or disability [1].
• This page covers how surgical errors happen, the factors that contribute to them, the WHO surgical safety checklist and the evidence behind it, never events, consent and open disclosure, and the audit and quality-improvement machinery that surrounds all of it.
• The recurring finding is that inadequate communication ranks highest in frequency among the factors contributing to patient safety incidents [1].

Definition

• A patient safety incident is a preventable event or circumstance that did or could result in unnecessary harm to a patient [1].
• The category has three parts: adverse events that result in actual harm, near-miss events that by chance or intervention cause no harm, and no-harm events that reach the patient but cause no harm because of chance or other mitigating circumstance [1].
• These must be distinguished from the recognised complications and side effects that follow from the underlying condition, comorbidity or the treatment itself, most medical care entails some level of risk [1].

A never event, or serious reportable event, is a subset of serious incidents considered wholly preventable (a retained abdominal swab or instrument, for instance) where guidance providing strong systemic protective barriers, namely checklists, should have been implemented [1]. Each type has the potential to cause serious harm or death, but serious harm or death does not have to have occurred for an incident to be categorised as a never event [1].

Quality improvement in health care is the continuous and combined efforts of people to make changes that lead to better patient outcomes, enhanced healthcare system performance and better learning and professional development [1].

Clinical audit, a function of clinical governance, is the means by which the care being provided is compared with accepted standards; it lets providers and patients know how a service is doing (quality assurance) and identifies where there could be improvements [1].

How errors happen

The problem of error can be viewed in two ways, from a person approach or from a system approach [1].

The person approach starts from the principle that humans are fallible, and that errors occur through doing the wrong thing (errors of commission; failing to act) errors of omission; or doing the right thing incorrectly, errors of execution [1]. Its limitation is that for most errors it tends to blame the individual and restricts learning, which is why individual performance is better assessed by professional appraisal and clinical audit, benchmarked against institutional, national and international norms [1].

The system approach recognises that adverse events rarely have an isolated cause, and that they are best addressed by examining why the system failed rather than who made the mistake [1]. The majority of near misses and adverse events are due to system factors [1].

Two models underpin this. Heinrich's safety pyramid proposes that each major injury within a system masks a multiple of minor injuries and near misses, which is the argument for near-miss reporting [1]. James Reason's Swiss cheese model explains harm as the consequence of multiple errors, analogous to holes in a Swiss cheese which, if they align, create a defect with adverse consequences; the corollary is that the more layers of responsibility, the fewer the chances of an adverse event [1].

• Safety-I and Safety-II describe two complementary stances.
• Safety-I places the emphasis on identifying errors after the event and preventing recurrence; Safety-II acknowledges that healthcare work is resilient and that everyday performance succeeds far more often than it fails, because clinicians constantly adjust what they do to match the conditions [1].
• Proactive safety management focuses on how everyday performance usually succeeds rather than why it occasionally fails, and actively strives to improve the former rather than simply preventing the latter [1].
• Sophisticated systems assimilate both [1].

The scale of the problem and the science of safety in Schwartz's account

• Medical error, counting system errors, is the third leading cause of death in the United States, about 251,000 deaths a year against 611,000 from heart disease and 585,000 from cancer in 2013, and the strategy is to make errors less frequent by designing for human limitations, more visible so they can be intercepted, and remediable when they reach the patient [2].
• Unnecessary care is itself error: it wastes an estimated $210 billion a year, over 2000 US physicians judged 21% of care unnecessary (22% of prescriptions, 25% of tests, 11% of procedures), the National Academy of Medicine puts a third of spending as wasteful, and opioid over-prescribing is the surgical exemplar, 249 million prescriptions in 2015 and 14 billion pills produced in 2016 (40 per citizen), surgeons being the commonest prescribers after pain specialists, so judicious discharge prescribing with non-opioid regimens is critical [2].
• High reliability organisations, the US Navy nuclear submarine programme, carrier flight decks, nuclear power plants, air traffic control, keep error rates far below what their risk predicts through individual ownership, attention to detail, mutual respect and free upward flow of information about mistakes; normal accident theory holds that in Perrow's high-risk systems (catastrophic potential, complex interdependent subsystems, tight coupling) accidents are intrinsic and must be treated as system problems, since "best people can make the worst errors as a result of latent conditions"; Donabedian's model measures structure (do the right tools, equipment and policies exist?), process (are they used?) and outcome (how often are patients harmed?), with culture (the collective attitudes and beliefs of caregivers) now the fourth measurable component linked to clinical outcomes [2].
• Schwartz's error taxonomy: adverse event (injury from management rather than disease, prolonging stay or disabling, preventable or not), negligence (care below what a reasonable similarly trained physician would give), near miss (error without harm, revealing system failures before harm) and sentinel event (unexpected death or serious injury, ABO-incompatible transfusion, wrong-site, wrong-procedure or wrong-patient surgery, fatal medication error, retained foreign body, demanding immediate investigation) [2].

Contributing factors

Bailey & Love groups the contributing factors into three families [1].

Human factors are inadequate patient assessment with delays or errors in diagnosis; failure to use or interpret appropriate tests; error in the performance of an operation, treatment or test; inadequate monitoring or follow-up; deficiencies in training or experience; fatigue, overwork and time pressure; personal or psychological factors such as depression or drug misuse; variation in the patient or working environment; and lack of recognition of the dangers of medical error [1].

System failures are poor communication between healthcare providers; inadequate staffing levels; disconnected reporting systems or over-reliance on automated ones; lack of coordination at handover; drug similarities; environment design and infrastructure; equipment failure from lack of parts or skilled operators; cost-cutting; and poor governance with inadequate systems to report and review incidents [1].

Medical complexity covers advanced and new technologies, potent drugs with their side effects and interactions, and the working environments of intensive care and the operating theatre [1].

• Fatigue is a genuine occupational safety risk, not a rite of passage.
• There is good evidence linking tiredness with medical error, and fatigue also causes depression, anxiety and confusion, all of which degrade performance [1].
• Burnout (a state of emotional, mental and physical exhaustion caused by prolonged stress) has increased in the medical workforce, attributed to loss of autonomy and increased administrative workload compounded by electronic record-keeping and a more regulated environment; doctors suffer burnout to a greater extent than other professions [1].
• Whether reducing trainee hours actually improves patient safety remains unsettled, because of the trade-off in requiring additional handovers and the consequent loss of continuity of care [1].

Adverse events and near misses go unreported for identifiable reasons (fear of blame and the potential for litigation among them) which is why clinical risk management must be carried out in a blame-free environment with an emphasis on education [1].

Culture, communication and the four sentinel cases in Schwartz's account

• Effective safety cultures acknowledge error-prone work, report without fear of punishment, collaborate across ranks and fund safety; traditional surgical culture opposes each, surgeons under-acknowledge fallibility, minimise stress effects and preside over a hierarchy in which intimidation once stopped nurses speaking up, ICU nurses report unresolved disagreement and exclusion from decisions, autonomy trumps cooperation, and safety stays a priority rather than a value until a high-profile event forces investment [2].
• Four cases shaped the field: Libby Zion (New York Hospital, 1984; missed pethidine interaction, physician fatigue) led the Bell Commission to cap residents at 80 hours a week and 24-hour shifts with a senior physician always present, adopted by New York in 1989 and by the ACGME in 2003; Betsy Lehman (Dana-Farber, 1994; fourfold chemotherapy overdose, a second patient harmed two days later) led to discipline of the doctor, three pharmacists and 15 nurses, an $11 million safety overhaul, full disclosure and a patient committee; Josie King (Johns Hopkins, 2001; methadone given against the mother's protest after a line was removed and fluids withheld, dying of dehydration) exposed sign-out failure and funded the Josie King Patient Safety Program; Mike Hurewitz (Mount Sinai, 2002; a living liver donor left to one first-year resident covering 34 patients, dying of aspiration) closed the transplant programme for over a year and produced New York's first living-donor guidelines [2].
• Communication breakdown is among the top three root causes of Joint Commission sentinel events and contributed to over 60% in 2011; one field study found communication failure in 30% of operating room exchanges, 36% of them substantively affecting safety; hand-offs, counts during critical steps and surgeon–pathologist messages relayed through intermediaries all degrade information; and perceptions diverge, 85% of surgeons rated collaboration with nurses high while only 48% of nurses rated surgeons so (anaesthetists rating each other 96%), nurses defining collaboration as having input respected and physicians as nurses anticipating and following instructions [2].
• The Safety Attitudes Questionnaire, adapted from aviation's cockpit management questionnaires and validated in over 500 hospitals, measures teamwork climate, safety climate, job satisfaction, perception of management, stress recognition and working conditions with items such as "I would feel safe being treated here as a patient" and "it is not difficult to discuss mistakes", benchmarks units and worker types and tracks interventions [2].

Errors in surgical practice

Coal-face errors, as described by Cuschieri and others, are those a surgeon can personally commit during the care of a patient: diagnostic and management errors, resuscitation errors, prophylaxis errors, prescription and parenteral administration errors, situation awareness, identification and teamwork errors, and technical and operative errors [1].

The recognised theatre errors are a short and specific list: the wrong patient in theatre, surgery on the wrong side or site, the wrong procedure performed, failure to communicate a change in the patient's condition, disagreement about proceeding, and retained instruments or swabs [1]. All are catastrophic for the patient and almost invariably occur through a lack of communication [1].

Failures of operative technique divide into cognitive errors of judgement, such as late conversion of a difficult laparoscopic procedure to open; procedural errors, where steps are not followed or are omitted; executional errors, such as using too much force; misinterpretation of anatomy or pathology, compounded in minimal access surgery by the limitations of a two-dimensional image; misuse of instrumentation, as with energised dissection; and missed iatrogenic injury, either at the time or diagnosed late [1].

• Proficiency is an acquired state, honed by teaching, practice and repetition, in which a surgeon consistently performs operations with good outcomes [1].
• In cognitive terms high proficiency is automatic unconscious processing (effortless, intuitive and untiring) as against non-proficient execution, which requires conscious control, constant attention and slow deliberate movement, and induces fatigue [1].
• The transition between the two is the learning curve, and it carries no negative connotation: a trainee at the conscious processing stage may still perform a perfectly good operation, though it will take longer and be more tiring [1].

Medication errors are among the most common errors across all specialties, because accuracy requires every step of ordering, dispensing and administration to be executed correctly [1]. The common ones are poor assessment or inadequate knowledge of the patient and their condition, inadequate knowledge of the medication, dosage calculation errors, illegible handwriting, and confusion or mixing up of drug names [1].

Never events, retained items and wrong-site surgery in Schwartz's figures

• National Quality Forum "never events" are unambiguous, usually preventable, serious (death, loss of body part or function) and adverse or indicative of system failure, in surgery, wrong body part, wrong patient, wrong procedure, unintended retained object, and intraoperative or immediately postoperative death in an ASA class 1 patient; 9744 paid malpractice claims for never events between 1990 and 2010 cost $1.3 billion with death in 6.6%, permanent injury in 33% and temporary in 59%, and 12.4% of named physicians were named again; the Joint Commission found wrong-site surgery occurring 40 times a week nationally [2].
• Retained surgical items occur in about 1 per 8000–18,000 operations (about 1500 a year in the United States, at least one per large hospital, an underestimate from claims data), more often in emergency surgery, unplanned changes of procedure, higher body mass index, multiple surgeons or procedures, staff changes and shifts spanning the case; sponges are commonest and present incidentally, as a mass or pain, or with abscess, erosion, fistula, obstruction, haematuria or a tumour-like lesion, and must always be removed; needles are usually incidental, rarely injurious, undetectable on plain film under 13 mm and may be left; instruments (malleable and "fish" retractors) present as pain or fullness; CT usually suffices; prevention rests on standard counts, wound exploration before closure, X-ray-detectable items only, debriefing after every case and radiography whenever any team member doubts the count or a risk factor exists, yet "falsely correct" counts occur in 21–100% of retained-item cases, counts distract nurses during critical steps, and items are retained despite known miscounts when surgeons dismiss them, so mandatory in-theatre radiography for any miscount removes conflict; legally the mere presence of the item proves improper care, though honesty and bedside manner mitigate [2].
• Wrong-site surgery (wrong patient, part, side or level) has no standard definition and is under-reported; estimates range from 1 in 112,994 to 1 in 15,500, Washington University suggests 1 in 17,000 (about 4000 a year, the third most frequent life-threatening error); of 427 Pennsylvania reports over 30 months more than 40% reached the patient and nearly 20% completed a wrong-site procedure; risks are multiple surgeons or procedures, time pressure, emergency, abnormal anatomy, obesity, incomplete assessment, weak verification and a culture revering the surgeon; communication errors underlie over 70%; surgeons operating on symmetric structures face a one-in-four career chance; orthopaedics and podiatry report 41%, general surgery 20%, neurosurgery 14%, urology 11%, with lower limbs 30%, head and neck 24% and genitourinary or groin 21% of sites [2].
• The Joint Commission's 1998 alert, 2003 National Patient Safety Goal and 2004 Universal Protocol (endorsed by over 50 organisations) mandate identity verification, site marking, a preoperative checklist, confirmation of documents and imaging, a time-out before incision with oral verification of patient, site, procedure, medications, records, imaging and equipment, and compliance monitoring; yet in Pennsylvania 21 wrong-site errors (12 completed) occurred despite proper time-outs and 16 despite correct marking, two-thirds of claims were preventable by verification protocols, and reported sentinel events have not fallen since 2004, perhaps through better reporting [2].
• Robotic surgery: the MAUDE database recorded over 10,000 device-related events in 14 years (98% from manufacturers, 2% voluntary), 1535 (14.4%) with significant harm (1391 injuries and 144 deaths) device failures making up about half, reports rising 32-fold since 2006 against a tenfold rise in cases, with disproportionate morbidity as new specialties adopt it, so validated curricula and simulation are needed as they were after laparoscopic bile duct injuries [2].

Strategies for patient safety

• The surgical safety checklist is the intervention with the best evidence behind it.
• Use of a perioperative checklist in eight hospitals around the world was associated with a fall in perioperative mortality from 1.5% to 0.7%, and in major inpatient complications from 11.0% to 7.0% [1].
• A later study from two Norwegian hospitals in 2015 showed complications falling from 19.9% to 11.5%, mean length of stay falling by 2 days, and hospital mortality falling from 1.9% to 0.2% in one of the two hospitals [1].

The checklist specifies checks at three obligatory time points, and the items are not intended to be comprehensive, additions and modifications to fit local practice are encouraged, as happened when COVID-19-specific checks were added during the pandemic [1].

• A checklist is not a substitute for a safety culture.
• Checklists are suited to some kinds of problem and not others, and managing patients involves a great deal of coordinated, time-pressured decision making; they are simple reminders, and unless coupled with attitude change and the removal of barriers to actually using them their impact is limited [1].
• If one begins to believe that safety is simple and all it requires is a checklist, there is a danger of abandoning other efforts to achieve safer, higher-quality care [1].

Successful implementation needs early engagement of staff, active leadership with local champions, extensive discussion, education and training, multidisciplinary involvement, coaching, ongoing feedback and local adaptation [1].

The theatre is not where most surgical errors happen. Several studies have shown that 53 to 70% of surgical errors occur outside the operating theatre, before or after surgery, and that greater improvement in safety comes from targeting the entire surgical pathway [1].

Briefings, debriefings, sign-outs and CUSP in Schwartz's account

• The five-point operating room briefing, names and roles, correct patient and procedure confirmed (the Universal Protocol time-out), antibiotics given, critical steps, potential problems, raises awareness of wrong-site errors, reports equipment problems early, cuts costs and delays (30.9% of staff reported delays before briefings, 23.3% after) and embeds antibiotic and DVT prophylaxis; debriefings treat errors as learning, designate follow-up, verify counts and read specimen labels aloud, since identification errors occurred in 4.3 per 1000 specimens (about 182 mislabelled a year in one hospital, unlabelled, empty containers, wrong laterality, site, or name) with consequences from repeat biopsy to wrong-site therapy; the WHO Surgical Safety Checklist structures sign-in before induction (identity, site, consent, marking, anaesthesia check, oximeter, allergy, airway and aspiration risk, blood loss over 500 mL or 7 mL/kg), time-out before incision (introductions, confirmation, antibiotics within 60 minutes, anticipated critical events, sterility, imaging) and sign-out before leaving (procedure name, counts, labelling, equipment problems, recovery concerns) [2].
• Sign-outs are variable and error-prone through omissions and illegible notes; the Navy's SBAR (situation, background, assessment, recommendation) structures them and every sign-out should open "In this patient, I am most concerned about…" [2].
• The Comprehensive Unit-based Safety Program measures culture first, engages management, holds regular multidisciplinary meetings where any staff member voices hazards, and applies local wisdom; in colorectal surgery it cut SSI by 33% through standardised skin preparation and pre-anaesthetic warming [2].
• Schwartz's best practices: conduct the time-out, brief, promote speaking up, screen high-risk cases radiographically for foreign bodies, and open sign-outs with the most likely immediate hazard [2].

Communication, consent and open disclosure

Poor communication is a common reason for patients taking legal action, and there are better treatment outcomes and fewer errors when communication is good [1].

Consent requires six pieces of information: the condition and why it warrants surgery; the type of surgery proposed and how it might correct the condition; the anticipated prognosis and expected side effects; the unexpected hazards; any alternative and potentially successful treatments; and the consequences of no treatment at all [1]. Patients should make these decisions without coercion or manipulation [1].

Consent should be obtained by someone capable of performing the surgery, and taken when the patient is fully aware, especially in the non-urgent situation, well before the procedure [1]. The frequent causes of an unsatisfactory consent process are failure to allow adequate time for discussion, and failure to understand that consent is a process which often requires multiple interactions with a patient [1].

Open disclosure is honest communication with a patient after an adverse event: a full explanation of what happened, the potential consequences, and what will be done to fix the problem [1]. Safe care also means looking after the patient after the event, ensuring the problem does not happen again, and sincerely offering regret or an apology as appropriate [1].

• Situational awareness and emotional intelligence are named as specific competencies rather than personality traits.
• Situational awareness describes an awareness of all individuals within the environment and an appreciation of the importance of change with time; emotional intelligence, as defined by Salovey and Mayer, is the ability to monitor one's own and other people's emotions, discriminate between them, label them appropriately, and use that information to guide thinking and behaviour [1].
• Both matter for identifying stress in oneself and in other members of the team [1].

Complaints are a source of learning. A complaint from a patient or carer often highlights a problem which, when analysed, provides an opportunity to reduce adverse events; patient advocacy units now provide options for resolving them, including information, mediation and conciliation meetings [1].

Risk management and disclosure in Schwartz's account

Half to two-thirds of hospital adverse events are surgical, mostly technical and in theatre, yet three-quarters of technical errors in claims involved fully trained surgeons within their expertise and 84% occurred in routine operations (complicated by comorbidity, anatomy, reoperation or equipment) so regionalisation and privileging limits may not reduce them; in the American College of Surgeons Closed Claims Study organ injury occurred in 40% but technical misadventure was the most deficient element in only 12%, communication and practice-pattern violations dominating in a third who received standard surgical care; common suits involve positional nerve injury, bile duct injury, missed or delayed diagnosis, wrong or delayed treatment, poor documentation, inappropriate indication, failure to consult and amputation [2]. Tone matters more than content: a "dominant" voice (deep, loud, fast, unaccented) signals lack of empathy and general and orthopaedic surgeons who sounded dominant were more often sued, while audible concern conveys empathy; rapport is the most important determinant of whether a suit is filed; errors must be disclosed immediately as an ethical and professional duty (concealment undermines trust and creates fraud liability), and disclosure improves rapport, satisfaction and claims, the Lexington VA's 1987 apology-and-disclosure programme moved it from among the highest claims totals to the lowest quartile over 7 years with average payouts of $16,000 against the VA's $98,000 and two trials in a decade, prompting VA-wide adoption in 2005, and the University of Michigan halved pending suits and cut litigation cost per case from $65,000 to $35,000, saving about $2 million a year [2].

Clinical governance, audit and quality improvement

Clinical audit compares care against accepted standards, and can be run nationally or locally within hospitals, departments, general practices or anywhere care is provided [1]. High-value quality improvement or audit activity has four stages: preparation and planning, measurement of performance, implementation of change, and sustainment and evaluation of the improvement [1].

Measuring clinical outcomes as part of a quality improvement cycle improves the quality of clinical care through shorter stays, better outcomes, fewer complications, reduced readmissions and greater patient satisfaction; informs national clinical audits; supports shared decision making; improves oversight of clinicians and teams; helps specialty associations become more transparent and patient-focused; supports appraisal and revalidation; and spreads best practice [1].

Comprehensive risk management is not an exercise in litigation avoidance but aims to develop cultural awareness and support for all healthcare workers in defining and delivering high-quality care [1]. Data collected from incidents should be collated and learnt from both institutionally and by uploading to a national database, and doctors should be familiar with the systems operating in their own working environment [1].

Patient-centred co-design is the process by which providers work in partnership with the people receiving care to identify and prioritise desirable outcomes, the speed of access to reliable advice, the effectiveness of treatment delivered by trusted professionals, the continuity of care and smooth transitions, involvement of and support for family and carers, availability of clear comprehensible information and support for self-care, and involvement in decisions [1].

Quality measures

• Without measurement it is impossible to determine whether a change actually improved quality [1].
• Measurement for improvement uses different methods from research, being concerned more with how best to introduce and replicate best practice than with determining for the first time what best practice should be [1].
• Quality is measured in three domains: structure, process and outcome [1].

Structural measures describe characteristics of the health system that affect its ability to meet the needs of patients or a population, usually the availability of material, infrastructural or human resources, the number of surgeons per 100,000 population, or the number of staffed theatre sessions in a hospital [1]. They are especially useful for evaluating and improving equity of access [1].

• Process measures assess what the provider did for the patient and how well it was done [1].
• Measurement for improvement most commonly tracks processes at the same site over time to reduce inappropriate variation, and process measures should be associated with better outcomes and ideally matter from the patient's perspective [1].
• The worked example is the start time of operating lists: starting the operating day on time reduces delays for patients awaiting surgery, shortens preoperative fasting and improves the ability to plan the theatre day [1].

• Outcome measures describe the effects of care on the health status of patients and populations, specific, observable, measurable changes representing the achievement of an outcome [1].
• Ideally they should be outcomes that matter to patients rather than to the provider, which is what a patient-reported outcome measure captures, and there should be evidence that they reflect the quality of the intervention [1].
• Outcome measures are what clinical audit commonly uses, comparing achieved outcomes against evidence-based standards [1].

National programmes and measures in Schwartz's account

• Practice-pattern or appropriateness measures are physician-defined and peer-benchmarked: the American College of Mohs Surgeons' mean tissue blocks per skin cancer averaged 1.7 nationally, outliers taking four or more were written to and offered coaching, and the approach extends to minimally invasive surgery utilisation and physiotherapy before elective spine surgery [2].
• AHRQ (founded 1989, 80% of its budget in grants) issued 27 Patient Safety Indicators from administrative data in 2003 (revised 2010), provider-level items such as anaesthesia complications, death in low-mortality groups, pressure ulcer, failure to rescue, retained foreign body, iatrogenic pneumothorax, postoperative hip fracture, haemorrhage, metabolic derangement, respiratory failure, embolism, sepsis, dehiscence, accidental puncture, transfusion reaction and obstetric trauma, as screening indicators rather than definitive measures [2].
• The Surgical Care Improvement Project (2003) targets infection, venous thromboembolism and cardiac events because complications affect 6% of non-cardiac and over 30% of high-risk operations, adding 3–11 days of stay, $1398 (infection) to $18,310 (thromboembolism) and up to 69% loss of median survival; SSIs are 14–16% of hospital-acquired infections, 2–5% after clean extra-abdominal and up to 20% after intra-abdominal surgery, prevention saving $3152 and 7 days per infection; cardiac events affect 2–5% of non-cardiac and up to 34% of vascular patients with 40–70% mortality per infarction and nearly half preventable by β-blockade; DVT follows about 25% and PE 7% of unprophylaxed major operations and low-dose heparin halves fatal PE; process measures are antibiotic within an hour, correct selection and stopping within 24 hours (48 cardiac), 6 a.m. glucose control after cardiac surgery, appropriate hair removal, colorectal normothermia, VTE prophylaxis ordered and given within 24 hours either side, and continuation of β-blockers, with outcome measures of 30-day infection, embolism, infarction, mortality and readmission, potentially saving over 13,000 Medicare lives a year [2].
• NSQIP, created by the Veterans Health Administration in 1991, cut 30-day mortality 31% and morbidity 45% in its first decade, was piloted at 18 private hospitals in 2001–2004 and adopted by the American College of Surgeons in 2004, 82% of 118 participating hospitals reducing complications by 11% and mortality by 17% a year, now with over 400 hospitals and 1.5 million cases; its risk-adjusted observed-to-expected ratios teach that safety is inseparable from quality, that adverse outcomes reflect systems more than individual competence, and that peer comparison is essential [2].
• Leapfrog (2000; purchasers covering 37 million people; 1200 hospitals in 41 regions surveyed in 2011) rewards computerised order entry, 24-hour ICU physician staffing, NQF safe practices and evidence-based referral with annual volume thresholds, CABG ≥450 hospital/100 surgeon, PCI 400/75, AAA repair 50/22, aortic valve 120/22, pancreatic resection 11/2, oesophagectomy 13/2, bariatric surgery 100/20, and hospitals adopting perioperative β-blocker policies cut open AAA mortality an estimated 51% (61% for endovascular repair at volume-compliant hospitals, not significant), with results published at hospitalsafetygrade.org [2].
• The WHO's Safe Surgery Saves Lives (second Global Patient Safety Challenge, launched after the 2004 World Alliance) pairs the checklist with population metrics (surgeons, anaesthetists, nurses and theatres per capita, volumes, mortality) and hospital metrics including the surgical Apgar score from intraoperative hypotension, tachycardia and blood loss [2].
• Transparency works: Chassin's public reporting of New York CABG mortality (1–18% by hospital against a 2% standard) produced a 41% statewide decline as hospitals held CUSP-like meetings, the 18% hospital falling to 7% in 3 years and 1.7% later; the Society of Thoracic Surgeons rewards transparency; and HCAHPS, the AHRQ-developed, NQF-endorsed (2005) standardised patient-experience survey (courtesy, listening, explanation, cleanliness, quiet, help, pain management), is reported publicly and drives 25% of the value-based purchasing score that sets Medicare payment, though many public rankings outpace measurement science [2].

Outcomes

• The financial argument runs alongside the clinical one.
• The 2017 OECD report The Economics of Patient Safety found that iatrogenic patient harm is the 15th leading cause of the global disease burden and accounts for 15% of all OECD countries' hospital expenditure [1].
• Unsafe care produces prolonged hospitalisation, loss of income, disability and litigation costing many billions of dollars every year [1].

• In low- and middle-income countries the probability of a patient being harmed in hospital is greater, with more healthcare-associated infection and difficulty maintaining equipment for want of parts or skills [1].
• In some countries the proportion of injections given with reused, unsterilised syringes or needles is as high as 70%, and unsafe injections cause 1.3 million deaths each year, primarily through transmission of hepatitis viruses and HIV [1].
• The WHO checklist shows that many patient safety initiatives are not resource intensive but require attention to the details of process and care pathways in the local context [1].

• The surgeon is a second victim.
• The concept has been advanced to acknowledge the effect of adverse outcomes on surgeons' well-being, without denigrating the harm caused to the first victim, the patient [1].
• Supporting surgeons after adverse events is increasingly recognised as a responsibility of the whole surgical community, and comprehensive risk management systems will need to be patient-centric while also acknowledging and supporting the resilience of those working in the service [1].

More than 100 million people worldwide undergo surgical treatment every year [1].

Saved on this device only — sign in to keep it if you switch devices.