Article Text
PDF
Abstract
Background In 2005, the injury compensation legislation in New Zealand was reformed to extend coverage for medical injury removing both ‘error’ and ‘severity’ from eligibility criteria. This led to an increase in claiming and claims acceptance rate, thus enlarging the treatment injury claims database. This database provides an unusual ‘no-fault’ perspective of patient safety events.
Methods The authors analysed the first 4 years of primary care treatment injury claims data to identify the type, incidence, severity and cause of injury in primary care.
Results There were 6007 primary care treatment injury claims; 64% were accepted as treatment injuries. Most claims arose in general practice (62%), and most claimants were female (62%). Most claims were assessed as minor (83%), 12% major, 4% serious and 1% sentinel. Medication caused most injuries (38%) and most serious and sentinel injuries (60%). Dental treatment caused 16% of injuries; injections and vaccinations combined caused 10%; and venepuncture, cryotherapy and ear syringing combined caused 13.5% of injuries, mostly minor. ‘Delay in diagnosis’ caused few injuries overall (2%), but a disproportionate number of serious and sentinel injuries (16%) and deaths (50%). Spinal/ neck manipulation caused 2% of serious and sentinel injuries.
Conclusions New Zealand's no-fault treatment injury claims database provides information about primary care patient safety events from an unusual ‘no-fault’ perspective. This analysis reinforces previous research identifying medication as a high-risk primary care activity and further identifies other primary care activities (dental care, injections, venepuncture, cryotherapy and ear syringing) as carrying important risks for patient harm.
- Adverse event
- general practice
- health policy
- patient safety
- primary care
Statistics from Altmetric.com
Introduction
New Zealand has had a tax-payer-funded accident insurance scheme to compensate people suffering injury, irrespective of fault since 1974.1 Prior to legislative reforms in 2005, compensation for medical injury was an anomaly in this otherwise no-fault scheme. However, under the 2005 reforms, the definition of eligible injury was expanded, the pre-existing ‘error’ requirement removed, and the prior duty for the Accident Compensation Corporation (ACC) to report all findings of ‘medical error’ to the registration authorities and the Health and Disability Commissioner waived. These changes freed providers to supply information about injury and assist patients in lodging claims for compensation without fear of disciplinary reprisal, and are consistent with the requirements of a no-blame culture of openness and learning as advocated by patient safety experts.2
‘Treatment injury’ is now defined as:
personal injury suffered by a person seeking treatment or receiving treatment and caused by treatment; and not a necessary part, or ordinary consequence, of the treatment, taking into account all the circumstances of the treatment, including the person's underlying health condition at the time of the treatment; and the clinical knowledge at the time of the treatment. (s.32)3
The liberal injury eligibility criteria created under these no-fault reforms have led to more claims being lodged (and accepted) enlarging the treatment injury claims database. This database provides a distinctive no-fault perspective of patient safety events and presents new opportunities for learning to guide injury-prevention initiatives unavailable to researchers in tort jurisdictions.4–6 To date, there has been no systematic analysis of this ‘no-fault’ treatment injury claims dataset. In this study, we analysed primary care treatment injury claims data from the first 4 years following the 2005 legislative reforms to discover the type, incidence and severity of injury in primary care, and to identify the events in primary care with the most severe potential consequences.
Methods
A patient can lodge a treatment injury claim with any ACC registered provider by completing both a general injury claim form (ACC45) and a treatment injury claim form (ACC2152).7 Information about the patient, injury, event, treatment provider and lodging provider is collected. ACC uses the term ‘event,’ rather than ‘treatment,’ to include injuries caused by failure to diagnose or treat. All claims are assessed by ACC both for eligibility for compensation (accept/decline) and for ‘potential consequences’ (see Table 1). Under the definition of treatment injury, all accepted claims are assessed by ACC as injuries bearing a causal relationship to care provided. Injuries are classified using both ICD10 and Read Code classification systems. A claim may be declined and yet still be assessed as serious if, for example, there is a medication dispensing or administration error which did not result in any injury but which nevertheless raised concerns about a risk of harm to the public. No assessment of injury preventability is made by ACC.
Accident Compensation Corporation definitions of ‘potential consequences’
For this study, the ACC provided deidentified data for treatment injury claims arising in primary care in the first 4 years following the legislative reforms (1 July 2005 to 30 June 2009). The settings identified by ACC as primary care included general practice/family medicine clinics, dental clinics, physiotherapy rooms, chiropractic rooms, osteopath rooms, community pharmacies, laboratories, radiology rooms and rest homes. Claims arising from care provided in hospitals, in private specialist clinics or by maternity providers were excluded.
All claims were analysed to discover claims' acceptance rate and potential consequences, settings where claims arose, and the age and sex distribution of claimants. Accepted claims (treatment injuries) were analysed to identify the type, incidence, severity and cause of injury. Serious and sentinel injuries were analysed to identify the events that caused injury which resulted in, or had the potential to result in, ‘death or major permanent loss of function.’ Declined serious and sentinel claims were analysed to identify the events assessed by ACC as having the potential to cause ‘death or major permanent loss of function’ (even though in this instance no injury resulted).
The analytical approach to these data was mainly descriptive, as we aimed to determine the content of the dataset and its ability to inform providers about patient safety events in primary care. Most analyses focus on the event, rather than the injury, because injuries can be caused by different events (eg, stroke may be caused by both neck manipulation and oral contraceptive medication).
Results
Primary care treatment injury claims and claimants
A total of 17 355 treatment injury claims from all healthcare settings were registered with ACC in the four study years;8 6007 of these were identified as arising from care provided in primary care settings (35%).
Figure 1 shows the potential consequences of claims. Most primary care claims were minor (83%), as were most claims overall (68%) (Tapp D, personal communication, 2010).
Potential consequences of treatment injury claims.
ACC accepted 64% of the primary care treatment injury claims (3845) as being injuries bearing a causal relationship to care provided. Only 58% of minor claims were accepted (2885) compared with 96% of major claims (701), 84% of serious claims (204) and 95% of sentinel claims (55).
Most primary care claims arose in general practice settings (62%), dental clinics (22%), laboratories (4%) and physiotherapy rooms (4%). Sixty-two per cent of claimants were female. Figure 2 shows the age distribution of claimants. Only 14% of claimants (852) were 70 years or older.
Age distribution of claimants.
Primary care treatment injuries (accepted claims)
There were 179 different types of primary care treatment injury classified in the database, but only 10 types of injury accounted for 66% of all treatment injuries. The most common injury was ‘allergic/adverse drug reaction’ (1334; 35%), followed by wound infection (416; 11%), haematoma (231; 6%) and nerve damage (199; 5%). Most injuries were minor (75.0%); 18.2% were major, 5.3% serious and 1.5% sentinel.
Table 2 shows the events causing injuries and serious and sentinel injuries in primary care. Medication caused most injuries (1457; 37.9%), and most serious and sentinel injuries (155; 59.6%). Dental treatment caused 16.3% of all injuries, but only 3.6% of the serious and sentinel injuries. Vaccination caused 5% of all primary care treatment injuries (209). Most vaccination injuries were caused by the injection process (172; 82%). Injections and vaccinations combined caused 10.2% of all injuries and 8.8% of serious and sentinel injuries, mostly infection or nerve damage. Venepuncture, cryotherapy and ear syringing combined caused 13.5% of primary care injuries. Most of these injuries were minor (89.6%).
Events causing injury in primary care
Delay or failure to diagnose caused few injuries overall (2.0%) but a disproportionate number of serious and sentinel injuries (15.0%). Cancer was the most commonly missed diagnosis (16), followed by infection (nine) (viral pneumonia, necrotising fasciitis and meningitis), testicular torsion (five), cardiac conditions (four, including two myocardial infarctions) ectopic pregnancy (two), and one each of pulmonary embolism, giant cell arteritis and congenital glaucoma and dislocation of the hip.
Neck manipulation by both physiotherapists and chiropractors caused four serious or sentinel strokes in women (aged 33, 39, 52 and 62 years). Intrauterine contraceptive devices caused 21 injuries (12 uterine perforations and nine other infections). Inadequate nursing care in rest homes led to 15 decubitus ulcers, and falls during patient transfer resulted in five injuries.
Table 3 shows medication injuries. Antibiotics caused 59.8% of medication injuries (841), but only 28.2% of the serious and sentinel medication injuries (42). Steroids (oral and injected) caused 10.4% of medication injuries (146) and 14.8% of the serious and sentinel injuries (22); and anti-inflammatory drugs (NSAIDs) caused 10.4% of medication injuries (146) and 12.8% of serious and sentinel medication injuries (19).
Medications causing injury in primary care
Of the individual drugs causing serious and sentinel medication injuries, prednisone was most commonly implicated (14), followed by diclofenac (11), augmentin (11), warfarin (10), terbinafine (six), nitrofurantoin (five), erythromycin (five) and ciprofloxacin (five).
Twenty-six injuries ended in death. Most deaths related either to delay or failure to diagnose or treat (13; 50%), or to medication (nine; 34.6%). Warfarin was the most commonly implicated drug (three).
Declined serious and sentinel claims
Forty-three claims were declined but assessed as either serious (40) or sentinel (three). Most were associated with either pharmacy medication dispensing (16; 37.2%) or rest home medication administration (15; 34.9%).
Discussion
Patients who suffer treatment injury in New Zealand are eligible for compensation irrespective of negligence or injury severity.3 This is the first study to analyse the no-fault treatment injury claims dataset to identify threats to patient safety in primary care.
One-third of all treatment injury claims arose in primary care settings, mostly in general practice. Most primary care contacts in New Zealand are also in general practice.9 Most claimants were women, reflecting women's greater use of primary care services.10 Most primary care treatment injuries were minor, consistent with previous research findings,11 12 but nearly 1000 injuries were more severe, and 26 resulted in death confirming the importance of addressing patient safety concerns in primary care.13
Allergic reaction was the most common injury and medication the leading event precipitating claims. This finding is consistent with previous research findings and reflects both the importance of medication in the primary care treatment repertoire and its dangers.11 12 14–16 Most medication injuries were caused by antibiotics (60%), anti-inflammatory drugs (10%) and steroids (10%). This does not mirror prescribing rates: although most primary care prescriptions are for antibiotics (15%), anti-inflammatory drugs are only sixth (3.5%) and steroids 11th (2%).17 Warfarin was an outstanding cause of medication-related harm, despite lessons from research extending over more than a decade providing information on how to manage warfarin use more safely.18–20 The need for more practical translational research in this area is compelling. Injections, a known source of preventable harm, were also a common cause of injury in the dataset.21 We also found a disconcerting number of strokes in young women caused by neck manipulation (four).22
Interpretation of study findings is limited by the incomplete picture of injury provided by the dataset. Maternity claims were not included in the study sample, and yet much maternity care is provided in primary care settings. Also, it is likely that both underclaiming and selective claiming occurred, because of either a lack of awareness or a lack of motivation to claim.23 Older people, who are known to be especially vulnerable to medical injury, were under-represented in the dataset—most likely because of a lack of claiming.14 23 24 The incentive to claim may be driven by the cost of remedial treatment, and because hospital care is provided free in New Zealand, this incentive is present only when a treatment injury is managed in primary care (where patients usually have to pay a part-charge). Diagnosis, a well-recognised challenge for primary care, was under-represented in the dataset (2% of injuries). This anomaly is most likely due to underclaiming rather than the exceptional diagnostic skills of New Zealand providers.12 25–31 Increased awareness about eligibility is likely to lead to increased claiming, and so any trends in claims cannot be taken as indicative of trends in injury.
Interpretation is further limited by the lack of a denominator for many events, meaning that we cannot determine RR. Thus, while both neck manipulation and hormonal contraceptives caused four strokes in women, we cannot say which treatment is more dangerous.
Interpretation for injury-prevention purposes is further curtailed by the lack of information about injury preventability.32 It is likely that many of the medication injuries were largely unpreventable, occurring with the proper use of drugs. However, antibiotics caused most medication injuries, and as the overprescribing of antibiotics in primary care has been identified as a problem, this analysis would support calls to reduce overprescribing.33
This study provides an uncommon no-fault perspective of patient safety events in primary care. Study findings reinforce established knowledge and can be used to inform patient care. Both the incidence and severity of injuries indicate targets for improvement, but further research is needed to discover patterns of injury causation to better inform patient safety initiatives. Further research is also needed to discover the effect of the no-fault legislative reforms (if any) on provider willingness to be open about injury and the development of a ‘no-blame culture of openness and learning’ in New Zealand.
Acknowledgments
We gratefully acknowledge K Thomas, RNZCGP, and D Tapp and R Taylor, ACC, for their help in obtaining data and for their helpful comments on early drafts of the paper.
References
Footnotes
Funding University of Otago, Emily Johnston PhD Scholarship.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.