Trends in Pediatric Malpractice Claims 1987–2015: Results From the Periodic Survey of Fellows

Data were collected through the Periodic Survey, a nationally representative survey mailed to randomly selected, nonretired US members of the AAP. Seven surveys were fielded in 1987, 1990, 1992, 1995, 2001, 2007, and 2015. Response rates ranged from 42% in 2015 to 67% in 1992 with an overall average of 58%. The number of completed surveys varied between 629 and 1071, totaling 5731 for all surveys combined. The AAP Institutional Review Board approved each of the Periodic Surveys.

For surveys from 1995 on, we assessed nonresponse bias by comparing the analytic samples to their respective target samples using key demographic variables from the AAP member database for the year of the survey (age, sex, and geographic region). Data were not available for the surveys before 1995. For each available year, χ² tests compared respondents and nonrespondents on the basis of sex and region, and t tests compared the samples on the basis of age.

Three measures of experience with malpractice were reported by respondents. For surveys 1987–2015, we analyzed (1) whether they had ever been the subject of a malpractice lawsuit or claim and (2) the outcome of the most-recent malpractice lawsuit or claim. Respondents who reported ever having been the subject of a malpractice suit or claim were asked to select 1 of 5 outcomes to describe the most-recent malpractice suit or claim: (1) plaintiff lost in court, (2) plaintiff dropped the claim or suit, (3) claim or suit in progress, (4) out-of-court settlement, and (5) plaintiff won in court.

For surveys 1995–2015, we analyzed indemnity amount paid by malpractice insurance carriers. Respondents who reported out-of-court settlement or the plaintiff won in court as the suit or claim outcome were also asked to report the indemnity amount paid by the malpractice carrier. All reported amounts were adjusted to 2018 dollars.

Across the surveys, we identified physician sex, subspecialty practice, practice location, and time trend as 4 factors that may affect malpractice experience, according to previous research. In addition, we examined longevity of career and workload as 2 additional factors that may increase exposure to malpractice risks. The 6 factors are defined as follows.

For surveys 1987–2015, we analyzed the following:

1. survey year,

2. pediatrician sex, and

3. longevity of career, defined as years lapsed between the year respondents reported that they went into practice, excluding residency, and the survey year.

For surveys 1995–2015, we analyzed the following.

1. Subspecialty practice: Practice type was defined as self-reported spending of at least 50% of professional time in (a) general pediatrics, (b) a hospital-based pediatric subspecialty (critical care, emergency medicine, neonatology and/or perinatology, or hospital medicine), (c) a nonhospital-based pediatric subspecialty, (d) administration, or (e) other, which includes all nonpediatric or undetermined specialties. We chose this categorization to separate nonpediatricians, who made up a small percentage of AAP membership, and administrators as well as to attempt to split subspecialists into hypothesized high- and low-risk groups. Because the Periodic Survey respondents have consistently identified the hospital as the most common location giving rise to a suit or claim, we explored whether hospital-based subspecialties are at higher risk.¹⁰ 

2. Workload: respondent-reported average number of hours in direct patient care in a typical week.

3. Primary practice location: respondent reported as urban inner city, urban not inner city, suburban, or rural.

The percentage of pediatricians reporting that they had ever had a malpractice lawsuit or claim and the outcome of the most-recent lawsuit or claim were described for each survey year. The association of malpractice experiences with the 6 factors (survey year, longevity of career, physician sex, subspecialty practice, workload, and practice location) was measured in bivariate analyses. Multivariable analysis was used to examine the independent effect of the 6 factors on the likelihood of ever having a malpractice lawsuit or claim, with analysis being limited to 1995–2015, the years when the full set of factors were available.

Distribution of the third measure, indemnity amount, was described in aggregate for all years since 1995, the first year when indemnity amounts were collected by the survey.

All analyses excluded residents and were performed by using SPSS software (release 25.0.0.1; IBM SPSS Statistics, IBM Corporation).

For each survey conducted since 1995, we found no difference between respondents and nonrespondents regarding region of the country, and no differences between respondents and nonrespondents on age or sex were seen for 1995. In 2001 only, female physicians were more likely to respond (P < .05). In the most-recent 3 surveys (2001, 2007, and 2015), the respondents had a higher mean age (by 2.1 years in 2001 and 3.5 years in 2007 and 2015) than nonrespondents (P < .001).

Respondent age averaged 46 years across surveys, ranging from 44.7 in 1992 and 1995 to 48.9 in 2015 (Table 1). Reflecting the growth in women entering pediatrics, the proportions of female respondents increased steadily throughout the study period from 23% in 1987 to 62% in 2015. Career longevity varied with survey year, decreasing to its lowest average of 13.2 years in 1995 but rising to 17.1 in 2015. Close to two-thirds of overall respondents identified as general pediatricians, and approximately one-fifth identified as pediatric subspecialists who practiced in either hospital-based (10.2%) or nonhospital-based (9.5%) subspecialties. The rest included either administrators (0.7%) or all other, which included nonpediatric specialties or undetermined practice areas (14.7%). Average number of hours spent per week in direct patient care decreased 4.7 hours from 38.9 in 1995 to 34.2 in 2015. Suburban, non–inner-city urban, inner-city urban, and rural primary practice locations averaged 39%, 28%, 22%, and 11%, respectively. Both subspecialty practice and primary practice location distribution remained stable between 1995, the first year these data were collected, and 2015.

Across all 7 surveys, 27.8% of pediatricians (n = 1580) reported ever having been the subject of a malpractice lawsuit or claim (Table 2). Bivariate analyses indicate that in 2015, 21.4% of pediatricians reported ever having been the subject of a claim or lawsuit, a decrease from a peak of 33% (P < .01) in 1990 (Fig 1). Male physicians were more likely to have been subject to a claim or lawsuit than female physicians (35.5% vs 16.9%; P < .01), as were hospital-based pediatric subspecialists compared with nonhospital-based subspecialists and generalists (35.8%, 25.7%, and 23.8%, respectively; P < .01). Risk of claim or lawsuit also increased with longevity of career (14.2% for those with 12 years [median longevity] or less in practice versus 40.5% for those with >12 years in practice; P < .01) and work hours (21.7% for those spending <40 hours per week [median work hours] in direct patient care versus 30.0% for those spending >40 hours; P < .01). Statistically significant differences were not found among practice locations (Table 2).

In the multivariable model (Table 3), all variables had a statistically significant relationship to the likelihood of a malpractice lawsuit or claim. Holding all other variables constant, the odds of having been the subject of a claim or suit decreased over time by a factor (adjusted odds ratio [aOR]) of 0.96 every additional year since 1995 (P < .01). The likelihood increased with being male (aOR 1.70; P < .01), being in a hospital-based pediatric subspecialty (aOR 2.16; P < .01) compared with general pediatrics, having a longer career (aOR 1.10; P < .01 for each additional year in practice), and working longer hours (aOR 1.01; P < .01 for each additional work hour per week in direct patient care). Practice location (urban, inner city, suburban, or rural) was nonsignificant overall (P = .06). However, the odds of a claim or suit with rural practice locations decreased significantly compared with inner-city urban locations (aOR 0.65; P < .05).

There were 1580 respondents who reported having been the subject of a malpractice claim or lawsuit, and of those, 1513 (95.8%) provided information on the result of the most-recent action. Of these respondents, 45.8% reported a favorable outcome (36.6% reported the claim or suit was dropped by plaintiff and 9.2% plaintiff lost in court); 34.3% reported an unfavorable outcome (32.8% reported out-of-court settlement and 1.5% reported the plaintiff won in court). The remainder, 19.8%, reported a lawsuit was still in process (Table 4). Trends (Fig 2) indicate that the proportion reporting that the plaintiff lost in court in the most-recent claim or suit increased from 2.3% in 1987 to 12.6% in 2015 (P < .01; Table 4). The proportion of respondents reporting having lawsuits still in process decreased from 26.2% to 9.6% (P < .01). Changes in other outcomes were statistically nonsignificant (Table 4).

Only the most-recent 4 surveys, those conducted since 1995, collected indemnity amounts. On the basis of these 4 surveys, 875 of 3376 respondents (26%) reported ever being named as a defendant in a claim or suit. Of these 875 respondents, 296 (33.8%) reported settling out of court (n = 280) or losing the suit or claim in court (n = 16). For the 280 respondents who reported that a suit was settled out of court, 224 (80.0%) reported any indemnity amount. Of 16 cases in which the plaintiff won in court, 14 (87.5%) reported any indemnity amount. Two respondents reported indemnity amounts of $10 574 and $212 197, respectively, without answering the question on lawsuit outcome. On the basis of these combined 240 cases of indemnity payments, we found a right-skewed distribution (Fig 3) with a median of $128 000. Three-quarters of indemnity amounts were <$421 000, and mean indemnity ($451 000) was in a similar range as the 75th percentile ($420 000) because of infrequently reported payments >$500 000.

Our analysis of the Periodic Surveys fell into 3 broad categories: change in the malpractice claim rate and outcome over time, characteristics of physicians associated with malpractice claims, and amount of malpractice indemnification.

The results show a reduction in pediatricians reporting at least 1 lifetime malpractice claim or suit from 33% in 1990 to 21% in 2015, a decrease of 36%. This accords with a reduction in claims against low-risk specialty physicians reported in a recent study based on malpractice insurance data.³  No single cause for this decrease has been identified. Possible explanations include improved medical care (eg, advances in vaccination), increased emphasis on patient-centered care, greater focus on patient safety, and malpractice reform.

In contrast to this reduction in the rate of claims, the unfavorable outcome rate was stable over the study period at 34%. Physicians responding to our survey reported an unfavorable outcome rate greater than insurance company data showing plaintiff success rates of 22% and 28%.^5,6  This may reflect that our respondents were reporting whether they had ever had a claim, whereas other studies were only looking for claims during a discrete time period (12 and 20 years). Our rate of pediatricians losing at trial was 1.5%, which accords with both pediatric studies^5,11  and physicians in general.¹²  Our data showed an increase in favorable outcomes for malpractice claims and suits over time; however, this gain appears to be due to a reduction in unresolved cases because the rate of unfavorable outcomes held steady (Fig 2). Although our survey did not specifically examine dismissed suits, our rate of dropped claims or suits of 36.6% is similar to the dropped, dismissed, and/or withdrawn rate in a previous study.¹¹  It is smaller, however, than the dropped, dismissed, and/or withdrawn rate reported in another pediatric study⁵  and when compared with physicians in general.¹² 

To our knowledge, this is the first pediatric study examining the physician factors impacting risk of malpractice claims and lawsuits. Little has been previously reported about physician factors associated with malpractice claims; most of the existing literature is older, in some cases 30 years old, and use little or no multivariable analysis.

Our data show that male pediatricians were more likely to be the subject of a malpractice suit or claim. This accords with nonspecialty-specific data showing that male physicians are more frequently the subject of medicolegal action¹³  and claims.^14–16  A meta-analysis examining sex differences in physician communication showed that female physicians were more likely to have positive communication with patients than male physicians.¹⁷  Good physician-patient communication has been associated with a lower risk of malpractice litigation.^18–21  It is thus not surprising that female pediatricians would be less likely to be the subject of a claim or suit than their male counterparts. We also found that pediatricians who practice in hospital-based subspecialties are at greater risk for claims and lawsuits. Both sex and subspecialty-practice differences persisted when adjusted for hours worked, career longevity, and practice location. As would be expected from a greater workload, we found that hours worked and career longevity were both positively associated with risk for claims and lawsuits. Although both career longevity and work hours increase exposure to patient care and any medical liability associated with that care, studies have additionally linked long physician hours to medical errors and patient-safety issues.^22,23  In examining practice location, we found that pediatricians working in rural locations had a significant decrease in the rate of claims and suits compared with those in inner-city locations. One previous study found that family physicians in a nonurban setting were more likely to be subject to a malpractice claim, however.¹⁴ 

We found a median indemnification of $128 000 with no significant change over the study period. This value accords with one large study examining indemnification in pediatric cases⁵  but is smaller than another study.⁶  Our mean indemnification of $451 000 was >75th percentile of indemnification amounts, suggesting the impact of infrequent but large verdicts. Both mean and 75th-percentile indemnity amounts appeared to increase over time, but large variability rendered the increases statistically insignificant. Our mean of $451 000 concurs with a mean of $438 000 (inflation adjusted to 2018 dollars from $414 000 in 2014 dollars) reported in a comprehensive review of all pediatric paid claims between 1992 and 2014 from the NPDB⁴  and with an earlier study (1985–2005).⁵  A much larger mean was reported in a third, smaller study.³ 

Varied sources make a substantial difference for any analysis of malpractice claims. Each data source, including NPDB, insurance data, and survey results, has its strengths and weaknesses. Although the NPDB provides a robust source of indemnification information and is a great source for average paid claim amounts, it does not have data on unpaid claims. Thus, this resource cannot be used to establish a true rate of claims. Large insurance databases provide robust data for claims rates and success ratios but are subject to multiple biases because they represent a select pool. Studies of both insurance company databases and the NPDB have limited details on physicians and physician practice and thus provide minimal information on physician demographics and the likelihood of claims.

By restricting our definition of hospital-based subspecialty to neonatology and/or perinatology, critical care, emergency medicine, or hospital medicine, we likely underrepresented pediatricians working with inpatients because some general pediatricians and nonhospital-based subspecialists work in inpatient settings to varied extents. Our imprecise measure of inpatient care involvement may have reduced its power in predicting risks for lawsuits and claims to a degree.

The Periodic Survey recipient pool contained all AAP fellows, including nonpediatrician members. This group includes pediatric surgeons, anesthesiologists, and other subspecialists who would be expected to have different malpractice patterns than pediatricians.³  Although we were able to isolate the effects that are attributable to nonpediatricians by controlling for subspecialty practice in the multivariable model of lawsuit risks, we could not do so in the bivariate analyses. To assess whether the inclusion of nonpediatrician data skewed lawsuit outcomes and indemnity amounts, therefore, we repeated those bivariate analyses with pediatrician-only data and found the results to be comparable (Supplemental Fig 4 and Supplemental Table 5).

Our study, based on survey results, has the benefit of providing specific data about physician factors associated with the risk of claims. There are, however, limitations in our data. Nonresponse bias was observed in some years, with older and female pediatricians being more likely to respond. Those who were subject to a claim may have been more likely or less likely to answer the survey. Additional limitations include a decreasing response rate over the survey period and a lower response rate to the question on indemnity amounts by respondents who reported an unfavorable claim or lawsuit outcome.

Data on practice type, workload, primary practice location, and indemnity amounts were not collected before 1995, limiting their use in the analysis. Details regarding physician characteristics, such as workload and primary practice location, were collected over variable periods after the reported lawsuits or claims occurred, and these characteristics may have changed between the time of the reported lawsuit or claim and the completion of the survey. This could result in increased measurement error and reduce the power of our models.

Our data do not include trainees, residents, or physician and nonphysician providers who are not AAP members. Findings reported in this study may therefore not generalize to all pediatric care providers.

Finally, the relatively low rate of unsuccessful outcomes made our indemnification mean appear to skew because of outliers (ie, large verdicts). Only larger investigations will show whether this skew is real, but it is validating that our mean and median data accord with other large studies.

Between 1987 and 2015, there has been a significant reduction in the proportion of pediatricians reporting ever having had a malpractice claim. Despite this reduction, male pediatricians and those in hospital-based pediatric subspecialties were significantly more likely to be sued. Medical malpractice is a fluctuating environment because both the law and medical practice are continually changing. It is critical, therefore, to revisit periodically malpractice risk and its impact on pediatricians.

We thank Julie Kersten Ake, senior health policy analyst at the AAP, who was instrumental in the coordination of this project, as well as Karen Wilson, MD; Lynn Olson, PhD; and the AAP Department of Research for their review of the article. We also thank the members of the AAP Committee on Medical Liability and Risk Management for their collaboration.

AAP

American Academy of Pediatrics

aOR

adjusted odds ratio

NPDB

National Practitioner Data Bank