Preventable medical errors are a major cause of morbidity and mortality in hospitalised patients. Several retrospective chart reviews find that 1–4% of hospitalised patients experience preventable adverse events.1^–5 Various methods are used to identify and collect data on these adverse events, but direct patient input is generally lacking.

In most industries, customer opinion of the quality of the product or service provided is important, and considerable resources are spent obtaining and analysing customer input. Few incident-reporting systems include patients’ perspectives, and some studies actually question patients’ ability to evaluate technical aspects of the care provided.6^–8 Medical staff and patients differ in their perceptions about healthcare outcomes.9 Healthcare providers’ professional roles and disciplines affect their ability to recognise quality issues.

STUDY DESIGN

This study’s aims were to determine: (1) the validity of patient malpractice claims as assessed by independent physician review, (2) actual malpractice claims rates (claims frequency ÷ clinical volume) by hospital, specialty and procedure, and (3) differences between Swedish and other systems’ malpractice claims rates.

Sweden, like Norway, Denmark and Finland, operates a no-blame malpractice system that compensates patients for preventable injuries received as a result of medical error.10 A single malpractice carrier, owned by the county councils responsible for providing healthcare to their inhabitants, insures all patients against injury resulting from medical errors. The system is not “no-fault,” as compensation is provided only for preventable injuries determined to result from a wrong diagnosis or treatment, or that could have been avoided using a less-risky intervention. It is for the practitioners, however, “blame-free,” and no record of their identities is maintained, nor is information shared with regulatory agencies or sanctioning bodies.

Patients file a malpractice claim by completing a simple form, available from the treatment facility or the internet. There is no filing fee, and providers are encouraged to assist patients to file claims where an error might have caused patient injury. Upon receipt of a claim, the insurance company obtains the complete medical record and additional information from the treating physicians. These records are initially reviewed by claims adjusters, and then by medical specialists with expertise in the patient’s type of illness and treatment. When recommended by the reviewers, compensation is paid to reimburse income loss, unreimbursed medical expenses, as well as pain and suffering caused by the injury. Punitive damages are not awarded. Payments are designed to supplement the extensive coverage offered by Swedish social and medical care systems—for example universal medical care, sick leave, disability support, etc.

METHODS

We examined all inpatient malpractice claims submitted to the Swedish Patient Insurance for alleged injuries occurring from 1997 through 2004. Claims data included type and nature of injury, affected body part(s), diagnoses and procedure codes, as well as information on county, hospital, department, patient age and gender. Radiology, pathology, oral surgery and anaesthesiology specialties typically do not have inpatient discharges, so claims rates for these specialties could not be calculated. The malpractice claims rate for ophthalmology is probably not representative because of the increased likelihood that the inpatient admission resulted from an outpatient procedure complication.

The hospital discharge database for the same period was provided by the Swedish National Board of Health and Welfare, and included key variables regarding all patient discharges from all Swedish hospitals with three or more malpractice claims annually during the period of study. These hospitals accounted for 98.5% of all Swedish inpatient malpractice claims during the study period. The database is evaluated annually, with a focus on filing routines, classification of data and the accuracy of the hospitals’ reporting data. Internal data error checks using age, gender and specific procedures revealed negligible error rates. Data from this study are provided to all participating hospitals annually, providing additional verification of clinical volumes, claims, etc.

We merged claims and hospital discharge data sets and calculated claims rates (number of claims÷discharges). We produced times series and comparisons across sites and surgical procedures using SAS and Excel software.

RESULTS

During the years 1997–2004, a total of 23 364 claims resulted from inpatient care provided by 76 hospitals reporting 11 514 708 discharges (table 1). The overall claims rate was 0.20%. For the first 6 years of observation, the annual variation was small (range 0.21–0.23). For the last 2 years, the rates were lower (0.18% in 2003 and 0.14% in 2004), mainly because claims from those years are still being received and processed. Claims must be filed within 3 years of injury detection but, in certain circumstances, are accepted up to 10 years after treatment.

During the study period, 49.5% (range 48.1–52.6) of all claims were judged valid by physician reviewers and eligible for compensation. Among these claimants, there were 2.4% deaths, 7.2% had disability >15%, and 25.3% required >3 months of sick leave. The claims rate for individual hospitals varied substantially, although these rates are not adjusted for case severity or surgical volume (fig 1). The distribution of claims across specialty shows that surgical specialties (including obstetrics and gynaecology) are disproportionately represented. Approximately 89% of claims and 99% of discharges were classifiable as either surgical or non-surgical specialties. Surgical specialties accounted for 88% of the claims and 90% of compensation paid but only 46% of the hospital discharges (table 2). The claims rate for surgical specialties was 0.36%, compared with 0.04% for the medical specialties (table 3). Differences are also seen between procedures producing the highest numbers of claims compared with procedures with the highest claims rates (table 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Claims rates and rates of paid claims in 76 Swedish hospitals in 1997–2004. Error bars denote ±2SD.

DISCUSSION

In the Swedish patient injury malpractice system, 49.5% patient complaints regarding injury resulting from medical errors were deemed valid by physician reviewers. Results from a large US Veterans Administration study suggest that patient complaints regarding problems with the quality of their care provide information not otherwise captured by typical “incident reporting systems.”11 Only 4.2% of patient generated malpractice claims had also been captured by the incident reporting system, and the tort claim was more likely to result in payment if an incident report had document patient injury. This suggests a systematic under-reporting in the incident reporting system.

Current methods of assessing the incidence of adverse events or medical errors give only a partial picture and are subject to several biases.4 Lee and Domino list a number of limitations of claims analysis based on the American Society of Anesthesiologists (ASA) Closed Claims Project including the large discrepancy between events and claims, a bias toward more severe injuries, lack of denominator data, and geographic bias.12 The present study included denominator data and all claims for the entire country. Vincent et al also focus on litigation-based material and echo their concerns, noting that litigated claims are highly selective and represent only those claims for which a public record exists.13 In litigated cases, complete records are rarely available, and often, causality and/or the value of the injury were contested and ultimately settled by judges and juries on the basis of conflicting medical opinions. Swedish claims data are not from litigated cases, and not subject to these limitations.

Swedish claims rates varied greatly across counties, hospitals, departments (specialties), and procedures. Specialties’ and specific procedures’ malpractice risks are typically determined by claims frequency.14 However, actual claims rates give a different picture. In this study only four procedures, all spinal disc operations, had both high frequency and rate. Gawande, et al identified 15 surgical procedures that accounted for 58% of surgical adverse events in their analysis of Colorado/Utah hospital records; 60% were also found on Sweden’s list of high volume procedures and included hysterectomy, hip replacement, and spinal disc procedures.15 Across specialties, orthopedics, general surgery, and obstetrics and gynaecology respectively had the most claims and accounted for almost three-quarters of the total malpractice expenditures (73.5%) during the 8 year study period. When examined by claims rates, orthopedics remains the highest with 13 of the 20 highest claims rate procedures, including seven related to intervertebral disc procedures. Obstetrics and gynaecology and general surgery are replaced in ranking by cardiac surgery and neurosurgery respectively. Recognising institutions, departments or procedures with significantly higher-than-expected malpractice claims rates could identify needed improvements that claims volume alone would miss. Also, the large observed differences in malpractice rates by specialty and procedure suggest that surgical mix and volume should be taken into account when comparing hospitals’ and physicians malpractice rates and would argue against the use of a specific, single, threshold number of cases as indicative of professional incompetence.

Although claims rates would appear useful indicators of medical error, fear of litigation in fault-based tort systems may prevent disclosure of medical errors to the patient, thereby reducing the number of claims filed. No-fault systems, by contrast, do not place the responsibility of a medical error on an individual practitioner and may both reduce barriers to file for compensation and increase the probability that an error is disclosed. Davis et al analysed hospital records and compensation claims for medical injury for the same year and region of New Zealand.8 Slightly more than 2% of hospital admissions were associated with potentially compensable adverse events. Although the claims process was well targeted, few claims were filed, and even fewer were actually compensated. The ratio of successful claims to events potentially eligible for compensation was approximately 1:30. In a more recent study, only one in 25 patients who experienced injuries that were both serious and preventable in New Zealand’s “no blame” system filed a complaint.16 In their study, approximately 14% of claims were paid following determination of medical error, compared with 50% in the Swedish study.

Comparable, published US claims rate data, which would require both the number of claims and the clinical volume associated with them, are not generally available. However, the University HealthSystem Consortium’s annual Claims Survey of its member academic medical centres showed a mean claims rate (patient demand for payment and lawsuits divided by the number of discharges) of 0.31 (0.018)% (1 SE) over the fiscal years 1996–2006. Participation in this internal member survey varied, but typically included 25–30 institutions with a total of 700 000–1 000 000 discharges annually.17 The pay rate of claims and lawsuits was approximately 26% over that period of time.

The high claims and pay rates for surgical specialties, compared with the non-surgical specialties, is not well understood but is found in other systems as well.2 8 15 It is possible that surgical errors are simply more obvious than those associated with medical errors, both to the patient and to the provider, but Gawande et al15 suggest that other factors, including invasiveness, culture of “surgical cure,” poorer doctor–patient relationship and treatment complexity account for the observed differences.

More recent studies linked patient complaints to physician malpractice risk. Hickson et al showed that patient complaints, of which only a minority specifically alleged medical error, predict malpractice risk for individual physicians.18 Murff et al, from the same institution, showed that surgical admissions with any type of patient complaint were more likely to be associated with major surgical complications, especially if the complaint was associated with an institutional “incident report.”19 Again, only about 30% of the complaints were specifically related to the quality of care. Whether the higher rates of complaint generation from the surgical specialties are related to their higher malpractice claims rates is unknown.

STUDY LIMITATIONS

Only inpatient claims from a single country were analysed. Only case abstraction data were used for this study, although complete medical records were used by the claims reviewers to assess the validity of the patient claim. Actual medical error rates were unknown, so the sensitivity of patient complaints as a surrogate for actual medical malpractice rates is unknown.

CONCLUSION

The Swedish patient-injury insurance system, based on patient-generated, physician-reviewed complaints alleging medical error and resultant injury, provides reliable, unique information on medical malpractice and allows comparisons between regions, hospitals, procedures and patient groups. This quantitative malpractice risk information is being used to inform hospital managers, clinical leaders and policy makers about needed improvements in patient satisfaction, quality and patient safety. Systems to collect and analyse patient complaints regarding alleged medical malpractice should be developed and supported nationally, regardless of existing malpractice adjudication systems.