Do doctors prescribe antibiotics out of fear of malpractice?

• Published date: 01 June 2022
• Author: Sebastian Panthöfer
• Date: 01 June 2022
• DOI: http://doi.org/10.1111/jels.12316

ORIGINAL ARTICLE

Do doctors prescribe antibiotics out of fear of malpractice?

Sebastian Panthöfer

Charles River Associates, Bruxelles, Belgium

Correspondence

Sebastian Panthöfer, Charles River Associates, Avenue Louise 143/b12, 1050 Bruxelles, Belgium.

Email: spanthoefer@gmail.com

Funding information

Spanish Ministerio de Economía y Competitividad, Grant/Award Numbers: MDM2014-0431, ECO2014-00450

Abstract

This paper studies whether doctors prescribe antibiotics to protect themselves against

potential malpractice claims. Using data from the National Ambulatory Medical Care

Survey on a representative sample of doctor visits from 1993 to 2011, I find that doctors

are significantly less likely to prescribe antibiotics following tort reforms that reduce mal-

practice pressure. The changing prescribing patterns appear to have no adverse effects on

patient health outcomes over the same time period. Almost 100 million hospital records

from the Nationwide Inpatient Sample reveal little to no effects of tort reforms on hospi-

tal stays involving conditions that can potentially be avoided through the timely use of

antibiotics. Taken together, these findings suggest that malpractice pressure induces doc-

tors to prescribe antibiotics that are medically unnecessary, thereby contributing to the

rise in antibiotic resistance.

KEYWORDS

antibiotic resistance, defensive medicine, malpractice pressure, misuse of antibiotics, tort reform

INTRODUCTION

Doctors in the United States and elsewhere prescribe too many antibiotics.

According to recent estimates, 30% of antibiotics prescribed in US ambulatory

care are inappropriate (Fleming-Dutra et al., 2016). The widespread mis- and

overuse of antibiotics promotes the growth of antibiotic resistance, which is

a serious public health issue that many developed countries face today. An

estimated 2.8 million antibiotic-resistant infections occur annually in the

United States, resulting in more than 35,000 deaths (CDC, 2019). The economic

impact of antibiotic resistance, while already substantial, is bound to increase

DOI: 10.1111/jels.12316

©2022 Cornell Law School and Wiley Periodicals LLC.

340 J Empir Leg Stud. 2022;19:340–381.

wileyonlinelibrary.com/journal/jels

over time as more and more bacteria will become resistant in the future given

the current trajectory.

One potential factor contributing to the overuse of antibiotics in the

United States are malpractice concerns of physicians. More than 7% of US-

based physicians face a malpractice claim in any given year (Jena et al., 2011).

In response to the threat of being sued, doctors have been found to resort to

defensive medicine, which refers to the provision of care that is suboptimal but

which reduces the (perceived) malpractice risk of physicians. Frequent prescrib-

ing of antibiotics may in part be due to defensive medicine, as antibiotics pro-

vide protection against serious bacterial infections that, if left untreated, can

lead to malpractice claims. Physician surveys and anecdotal evidence support

this hypothesis,

1

but evidence from actual clinical encounters is lacking to date.

This paper is the first to study the effect of malpractice pressure on antibiotic

prescriptions based on visit-level data. Using restricted-use data from the

National Ambulatory Medical Care Survey (NAMCS), a nationally representa-

tive sample of visits to office-based physicians in the United States, I estimate

the causal effect of malpractice pressure on antibiotic prescription rates with a

difference-in-differences design based on the variation in tort reforms across US

states from 1993 to 2011. The dataset, which was created specifically for the pur-

pose of this analysis, contains detailed visit-level information as well as informa-

tion on drugs that can act as potential substitutes for antibiotics and which may

therefore also be affected.

Results show that doctors respond to malpractice pressure by prescribing

antibiotics more frequently. After the introduction of a cap on noneconomic

damages—a commonly adopted tort reform that reduces malpractice pressure—

doctors are 6.3% less likely to prescribe antibiotics to a given patient. This effect

is statistically and economically significant and robust to a variety of model

specifications, including specifications designed to (informally) test for legisla-

tive endogeneity. Extrapolated to the US population, it translates into 3.2 mil-

lion (3%) fewer ambulatory care visits per year in which doctors prescribe

antibiotics if the states 21 US states that currently do not put a cap on noneco-

nomic damages would adopt a cap.

To better understand the mechanisms driving this effect, I consider addi-

tional prescription outcomes and heterogeneous reform effects across subgroups

of patients and physicians. Results show that the change in antibiotic prescrip-

tion rates is primarily due to a change in the prescriptions of narrow-spectrum

antibiotics rather than broad-spectrum antibiotics. The latter act against a wider

range of bacteria, but they also generally cost more, are more likely to cause side

effects, and have a bigger impact on the development of antibiotic resistance.

1

For instance, 33% of the physicians who participated in a survey on medical malpractice reported that they

frequently prescribe more medication than medically indicated in response to malpractice pressure, and a further

36% reported that they occasionally prescribe more medication than medically indicated (Studdert et al., 2005).

DEFENSIVE ANTIBIOTICS 341

Results also show that doctors do not prescribe less medication overall after

noneconomic damages caps are adopted. However, regressions including poten-

tial substitutes for antibiotics suggest that no substitution occurs for the antibi-

otics that are no longer prescribed after noneconomic damages caps are

enacted. Regressions allowing for heterogeneous effects of tort reforms on anti-

biotic prescription rates reveal three factors that interact with malpractice pres-

sure, in line with economic incentives: patient age, patient race, and type of

health insurance.

As part of a wider set of informal tests for legislative endogeneity, I examine

the pre-reform antibiotic prescription rates of states that adopt and states that

do not adopt noneconomic damages cap reforms. These reveal an interesting

asymmetry, which has previously been found in the context of cesarean sections.

When noneconomic damages caps are enacted, the pre-reform antibiotic pre-

scription rates of the states that adopt the caps and the respective comparison

group are statistically indistinguishable. However, when noneconomic damages

caps are abolished, the states that repeal the caps exhibit a higher antibiotic pre-

scription rate in the 6 months leading up to the repeal than the states that do

not repeal a cap at the same time. As Currie and MacLeod (2008) have argued,

this asymmetry arises because law changes generally apply retroactively when

tort reforms are ruled unconstitutional, whereas the old law prevails until a new

tort reform is finally adopted. Therefore, if doctors anticipate that a noneco-

nomic damages cap will be abolished, it is rational for them to adapt their prac-

tice patterns ahead of the repeal. In the present case, these anticipation effects

lead to higher antibiotic prescription rates in the period in which the cap

still applies, which suggests that the estimates for the impact of noneconomic

damages caps on antibiotic prescription rates are conservative.

Having established that noneconomic damages caps reduce antibiotic pre-

scription rates, the next natural question is whether patients are adversely

affected by the change in prescribing patterns. To answer this question, I draw

on the Nationwide Inpatient Sample (NIS), which is the largest publicly avail-

able all-payer inpatient healthcare database in the United States. Using data on

almost a 100 million hospital stays from 1993 to 2011, I investigate whether

noneconomic damages cap reforms lead to a change in hospital stays for condi-

tions that can be linked to antibiotic use. With the possible exception of mas-

toiditis, I find no evidence suggesting that noneconomic damages caps increase

the incidence of several bacterial infections for which hospitalization can poten-

tially be avoided by timely antibiotic prescriptions in ambulatory care. Results

do, however, reveal a significant decrease in the share of hospital stays for

Clostridium difficile infections, which are to a large degree caused by antibiotic

use. These infections could potentially be avoided through more cautious

antibiotic prescribing.

Taken together, the empirical results suggest that malpractice pressure

induces physicians to defensively prescribe antibiotics that have no clear health

342 DEFENSIVE ANTIBIOTICS