When informed consent goes poorly: a descriptive study of medical negligence claims and patient complaints

We established definitions for claims, complaints and cases (Box 1). Our study definition of what constituted a dispute over informed consent (a “case”) was patient centred.8 We did not seek to evaluate the reasonableness or legitimacy of patients’ allegations, with one exception: if a patient alleged a failure to warn of the risks of an outcome that occurred, and there was strong clinical evidence in the case file that the outcome had not in fact occurred, we excluded the case.

Our sample frame consisted of all medical negligence claims brought against Avant-insured doctors in New South Wales, Victoria and Queensland between 1 January 2002 and 31 December 2008, and complaints lodged with the HSC during the same period.

From 7846 claims, we identified 397 candidate cases through searching for keywords (eg, informed, consent, warn, risks, explain, disclose) in free-text précis of the allegations (two to five sentences written by claim managers). We then examined the full claim file on each of these to determine whether it met our case definition (Box 2). About 20% of complaints received by the HSC cannot be resolved easily by the parties involved and are referred to a dispute resolution process. These “conciliated” complaints were our focus. After ruling out conciliated complaints in several categories (access, administration, cost) that were extremely unlikely to have involved consent issues, three physician reviewers examined the hard-copy files associated with the rest (1415) to determine which met our case definition (Box 2).

The data for this study came primarily from hard-copy files. The variables of interest were determined through a review of the informed consent literature and a pre-review of 20 claim files and 20 complaint files. We collected information on patient characteristics (sex, age); doctor characteristics (sex, age, trainee status, specialty); location of the incident (type of facility, clinical setting); the specific treatment involved in the episode of care; the patient’s allegations; and the case outcome.

Patients’ allegations regarding how the consent was deficient were coded into eight predefined categories derived from a review of the informed consent literature and refined in our pilot work. Reviewers were permitted to code up to five allegation types, but were directed to identify a primary allegation, defined as the one about which the patient was most concerned. This information typically came from the initial statement of claim or complaint letter, or subsequent allegations made by the patient during the litigation or conciliation process.

We developed an electronic data collection instrument to facilitate collection of these variables and conducted a pilot test of the instrument on 30 cases before the formal review began, making minor modifications as appropriate.

Two medically qualified reviewers conducted the reviews onsite at Avant offices (Melbourne, Sydney and Brisbane) and the HSC (Melbourne) between March 2009 and January 2010. The reviewers underwent a day-long training session that covered the structure and content of claim and complaint files, use of the data collection instrument, and confidentiality.

To test the reliability of the review, a random subsample of cases consisting of about 10% (27) of the claims and 10% (23) of the complaints was re-reviewed by the second reviewer, who was blinded to the first review. We report κ scores9 for the primary allegation type and all allegation types because they are the only variables that required implicit judgements by the reviewers.

Our analyses are descriptive. Counts and percentages for the variables of interest generally did not differ between claims and complaints. We therefore reported results in aggregated form, except for variables with intrinsic differences in structure (eg, case outcomes).

We calculated the frequency of informed consent disputes by specialty using rates and rate ratios. Complaint rates are the total number of consent-related complaints over the study period, divided by the sum of annual counts of registered doctors in Victoria over the study period (“doctor-years”).10 Complaint rate ratios were calculated by dividing the complaint rate for each specialty by the complaint rate for a reference group (general practitioners). Differences between rate ratios for specialties were calculated using x2 tests.

We used the same method to calculate rates and rate ratios for claims by specialty. However, Avant’s member numbers within each specialty are commercially sensitive and could be back-calculated from specialty-level claim rates. Therefore we present only rate ratios for claims.

All analyses were conducted using Stata SE, version 10.0 (Stata Corp, College Station, Tex, USA).

More than half (57%) of the cases were against surgeons (Box 4). Four surgical subspecialties — plastic, general, orthopaedic and ophthalmic surgery — accounted for 81% of all cases against surgeons. Obstetrics–gynaecology (14%) and general practice (11%) were the other prevalent specialties. Collectively, surgeons, obstetrician-gynaecologists and GPs were involved in 82% of cases.

The rate of complaints against plastic surgeons was significantly higher than that against any other type of specialist. Rate ratios indicated that this was true for claims as well.

A comparison of rate ratios between claims and complaints exposed some noteworthy differences. Plastic surgery, neurosurgery, orthopaedics and vascular surgery had the highest rate ratios for claims. By contrast, there were no informed consent complaints against neurosurgeons; and dermatology ranked second highest for complaints, compared with its ranking of ninth for claims.

In 71% of cases, the primary allegation concerned a complication of treatment that had not been mentioned or fully understood, and then materialised (Box 5). The next most common types of primary allegations were that the scope of the consent had been exceeded (10% of cases), the risk that the procedure would confer no benefit (as opposed to harm) had not been mentioned (6%), and the process by which consent was obtained was unsatisfactory (6%). Process allegations involved situations in which patients felt rushed, pressured to proceed, or regarded the language used as incomprehensible. No case had as its primary allegation that there were problems with consent to experimental treatments or research protocols.

Our instrument allowed the coding of secondary allegations, of which there were 205 (43% of cases; mean, 1.6 per case). Widening the analysis in this way showed increased dominance of undisclosed complications (78% of cases) as an allegation; it also increased the proportions of cases in which patients were dissatisfied with the consent process (to 17%) and were unaware of alternative treatment options (to 14%).

Reliability testing, based on 50 pairs of reviews, showed excellent agreement between reviewers. They agreed on the primary allegation type in 92% of the double reviews and overlapped on the allegations identified in 99% of cases. The κ scores were 0.81 (95% CI, 0.58–0.95) and 0.98 (95% CI, 0.92–1.00), respectively.

The disputes arose in relation to a wide array of treatments; however, clustering was evident. Thirty specific treatments accounted for 71% of cases (Box 6). There was a heavy orientation toward operations: 92% (442/481) of cases involved surgical procedures and 16% (77/481) involved cosmetic procedures. Five types of treatments — procedures involving reproductive organs (57/481; 12% of cases), procedures involving facial features excluding eyes (57/481; 12%), prescription medications (40/481; 8%), eye surgery (36/481; 7%) and breast surgery (32/481; 7%) — accounted for 46% of all cases.

The cases took an average of 1.8 years (interquartile range [IQR], 0.5–3.6 years) from filing to closure, with complaints (median, 1.0 years; IQR, 0.6–1.6 years) resolved more quickly than claims (median, 2.3 years; IQR, 1.3–3.5 years). Plaintiffs received payments in 51% (135/263) of claims. Eleven claims proceeded to court verdict, and plaintiffs won damages in three. Monetary restoration was paid to complainants in 27% (59/218) of complaints but, consistent with the functions of the HSC, a much wider range of remedies was sought and obtained in these and other complaints.