International classification of diseases codes showed modest sensitivity for detecting community-acquired pneumonia

doi:10.1016/j.jclinepi.2006.10.018

Journal of Clinical Epidemiology

Volume 60, Issue 8, August 2007, Pages 834-838

https://doi.org/10.1016/j.jclinepi.2006.10.018 Get rights and content

Abstract

Objective

To estimate the sensitivity of International Classification of Diseases (ICD-9-CM) coding for detecting hospitalized community-acquired pneumonia and to assess possible determinants for misclassification.

Study Design and Setting

Based on microbiological analysis data, 293 patients with a principal diagnosis of community-acquired pneumonia at seven hospitals in the Netherlands were assigned to three categories (pneumococcal pneumonia, pneumonia with other organism, or pneumonia with no organism specified). For these patients, the assigned principal and secondary ICD-9-CM codes in the hospital discharge record were retrieved and the corresponding sensitivity was calculated. Furthermore, pneumonia-related patient characteristics were compared between correctly and incorrectly coded subjects.

Results

The overall sensitivity was 72.4% for the principal code and 79.5% for combined principal and secondary codes. For pneumococcal pneumonia (ICD-9-CM code 481) and pneumonia with specified organism (ICD-9-CM code 482–483), the sensitivities were 35% and 18.3%, respectively. Patient characteristics were not significantly different between correctly and incorrectly coded subjects except for duration of hospital stay, which correlated negatively with coding sensitivity (P = 0.01).

Conclusion

ICD-9-CM codes showed modest sensitivity for detecting community-acquired pneumonia in hospital administrative databases, leaving at least one quarter of pneumonia cases undetected. Sensitivity decreased with longer duration of hospital stay.

Introduction

Community-acquired pneumonia (CAP) is a common and potentially fatal infection of lung tissue and is associated with high health care costs. Therefore, CAP is subject to many epidemiological and economical studies. In about 20% of all pneumonia cases, inpatient treatment is required because the clinical situation does not allow outpatient therapy [1], [2]. Many of the studies on pneumonia, therefore, identify cases of hospitalized CAP because these are most likely to result in death and resource use. A common approach to identify cases of hospitalized CAP is using hospital discharge records as coded according to the International Classification of Diseases-9th revision-Clinical Modification (ICD-9-CM). The common ICD-9-CM codes used for this purpose are 481, 482.x, 483.x, 485, and 486 [3], [4], [5], [6], [7], [8]. Information, however, on the validity of such data is inconclusive or lacking in many cases. Several researchers have assessed the validity of hospital discharge records of various diseases, including pneumonia, by identifying cases through ICD-9-CM codes and subsequently reviewing medical charts to confirm or reject the correct diagnosis [9], [10], [11], [12]. This approach, however, leaves cases of CAP without an ICD-9-CM code for pneumonia undetected and provides no information on the sensitivity of ICD-9-CM coding for detection of cases of CAP. The aim of this study is to estimate the sensitivity of ICD-9-CM code assignment in a population of patients admitted with a principal diagnosis of CAP. Furthermore, we aim to assess possible determinants for misclassification.

Section snippets

Patients and methods

This study used patient data from a randomized open label clinical trial (July 2000–March 2004) on efficacy of an early switch of intravenous antibacterial treatment to oral treatment of CAP [13]. All adult patients hospitalized for CAP in seven hospitals (two university medical centers and five teaching hospitals) in the Netherlands were eligible for inclusion in that study. Pneumonia was defined as a new or progressive infiltrate on a chest X-ray plus at least two of the following criteria:

Results

Of the 293 patients hospitalized for CAP, 40 (14%) had confirmed pneumococcal pneumonia, 82 (28%) had pneumonia with another organism specified, and 171 (58%) had pneumonia with no organism specified (Table 1). In total, 212 patients had any pneumonia-related ICD-9-CM code (481–486) as principal diagnosis, yielding an overall sensitivity for any pneumonia-related ICD-9-CM code of 72.4%. The overall sensitivity for six of the seven participating hospitals separately ranged from 61.5 to 82.0%

Discussion

Our study showed that in patients hospitalized with confirmed CAP, overall, only 72% was assigned any ICD-9-CM code for pneumonia (481–486) as the principal diagnosis on the hospital discharge record. For “pneumococcal pneumonia” and “pneumonia with other organism specified”, sensitivity was as low as 35% and 18.3%, respectively.

Ideally, the ICD-9-CM code in the primary position (principal diagnosis) on the hospital discharge record always represents the medical condition that is chiefly

Acknowledgments

We gratefully thank all hospitals that participated in this project.

References (18)

M.A. Woodhead et al.
Prospective study of the aetiology and outcome of pneumonia in the community
Lancet
(1987)
K.L. Movig et al.
Validity of hospital discharge International Classification of Diseases (ICD) codes for identifying patients with hyponatremia
J Clin Epidemiol
(2003)
M.L. De Bruin et al.
Hospital discharge diagnoses of ventricular arrhythmias and cardiac arrest were useful for epidemiologic research
J Clin Epidemiol
(2005)
M.C. McKean
Evidence based medicine: review of BTS guidelines for the management of community acquired pneumonia in adults
J Infect
(2002)
J. Segreti et al.
Principles of antibiotic treatment of community-acquired pneumonia in the outpatient setting
Am J Med
(2005)
P. Bi et al.
Trends in mortality rates for infectious and parasitic diseases in Australia: 1907–1997
Intern Med J
(2003)
E.M. Mortensen et al.
The effect of prior statin use on 30-day mortality for patients hospitalized with community-acquired pneumonia
Respir Res
(2005)
J. Almirall et al.
Risk factors for community-acquired pneumonia in adults: a population-based case-control study
Eur Respir J
(1999)
E.M. van de Garde et al.
Angiotensin-converting enzyme inhibitor use and pneumonia risk in a general population
Eur Respir J
(2006)

There are more references available in the full text version of this article.

Cited by (81)

Hospitalised adults with pneumonia are frequently misclassified as another diagnosis
2019, Respiratory Medicine
Using data from a large randomised controlled trial of adults hospitalised with acute respiratory illness, we examined the reliability of pneumonia diagnosis on discharge documentation. 50 (28.2%) of 177 patients with a pneumonia diagnosis had no radiological evidence of pneumonia. 67 (34.9%) of 192 patients with clinico-radiological evidence of pneumonia did not have a diagnosis of pneumonia listed; ‘COPD exacerbation’ or ‘lower respiratory tract infection’ was often listed instead. These patients more frequently had a respiratory comorbidity and lower oxygen saturations, CRP and temperature at presentation. Pneumonia diagnoses misclassification on discharge documentation may have clinical, financial, and research data implications.
Impact of infant pneumococcal conjugate vaccination on community acquired pneumonia hospitalization in all ages in the Netherlands
2017, Vaccine
The long-term impact of pneumococcal conjugate vaccines on pneumonia hospitalizations in all age-groups varies between countries. In the Netherlands, the 7-valent pneumococcal conjugate vaccine (PCV7) was implemented for newborns in 2006 and replaced by PCV10 in 2011. We assessed the impact of PCVs on community-acquired pneumonia (CAP) hospitalization rates in all age-groups.
A time series analysis using Poisson regression was performed on 155,994 CAP hospitalizations. Hospitalization rates were calculated using the total number of hospitalizations as denominator. The time trend in the pre-PCV period (1999–2006) was extrapolated to predict the hospitalization rate in the post-PCV period (2006–2014) if PCV had not been implemented. Rate ratios over time were calculated by comparing observed and predicted time trends.
In children <5 years of age, the observed hospitalization rates during the post-PCV period were significantly lower than predicted if PCV had not been implemented (0–6 months: 0.62, 95% CI: 0.41–0.96; 6 months – 1 year: 0.67, 95% CI: 0.50–0.90; 2–4 years: 0.78, 95% CI: 0.61–0.97). In all other age-groups, rate ratios declined over time but did not reach statistical significance.
After introduction of PCV, CAP hospitalizations declined in young children but no clear impact of PCV on CAP hospitalizations was seen in other age-groups.
Changes in pathogens and pneumococcal serotypes causing community-acquired pneumonia in The Netherlands
2017, Vaccine
Citation Excerpt :
To the best of our knowledge, there is only one study that reported an overall decrease in proportion of pneumococcal CAP in adults after introduction of childhood vaccination with PCV7 [16] . The latter study, however, identified patients based on ICD coded hospital discharge records, which have limited reliability regarding aetiology [17] . Studies using extensive microbiological diagnostics are needed to be able to assess this in more depth.
In 2006 a 7-valent pneumococcal conjugate vaccine (PCV7) was introduced in the immunisation programme for infants in The Netherlands and replaced by PCV10 in 2011. Limited data exist about the impact of PCV on the aetiology of CAP as a whole. The aim of the present study is to describe the overall changes in microbial aetiology, pneumococcal burden (including non-bacteraemic pneumococcal pneumonia) and its serotypes in adult community-acquired pneumonia (CAP) after the introduction of these PCVs.
Hospitalised adult CAP patients who participated in three consecutive trials were studied (2004–2006 (n = 201), 2007–2009 (n = 304) and 2012–2016 (n = 300) and considered as pre-PCV7, PCV7 and PCV10 period). Extensive conventional microbiological testing was applied for all patients. In addition, patients with a serotype-specific pneumococcal antibody response were diagnosed with pneumococcal CAP. Changes in proportions of causative pathogens and distributions of pneumococcal serotypes were calculated.
The proportion of pneumococcal CAP decreased from 37% (n = 74/201) to 26% (n = 77/300) comparing the pre-PCV7 period with the PCV10 period (p = 0.01). For other pathogens, including Legionella spp., Mycoplasma pneumoniae, S. aureus, H. influenzae, and respiratory viruses, no sustained shifts were observed in their relative contribution to the aetiology of CAP. Within the pneumococcal CAP patients, we observed a decrease in PCV7 and an increase in non-PCV10 serotype disease. PCV10-extra type disease did not decrease significantly comparing the PCV10 period with the pre-PCV7 and PCV7 period, respectively. Notably, PCV7 type disease decreased both in bacteraemic and non-bacteraemic patients.
Our findings confirm that PCV introduction in infants impact the microbial aetiology of adult CAP and suggest herd effects in adults with CAP after introduction of PCVs in children.
The value of electronically extracted data for auditing outpatient antimicrobial prescribing
2018, Infection Control and Hospital Epidemiology
Validation and incidence of community-acquired pneumonia in patients with type 2 diabetes in the BIFAP database
2017, Epidemiology and Infection
Best Practices for Identifying Hospitalized Lower Respiratory Tract Infections Using Administrative Data: A Systematic Literature Review of Validation Studies
2024, Infectious Diseases and Therapy

View all citing articles on Scopus

View full text

Original ArticleInternational classification of diseases codes showed modest sensitivity for detecting community-acquired pneumonia

Abstract

Objective

Study Design and Setting

Results

Conclusion

Introduction

Section snippets

Patients and methods

Results

Discussion

Acknowledgments

Lancet

J Clin Epidemiol

J Clin Epidemiol

J Infect

Principles of antibiotic treatment of community-acquired pneumonia in the outpatient setting

Am J Med

Trends in mortality rates for infectious and parasitic diseases in Australia: 1907–1997

Intern Med J

The effect of prior statin use on 30-day mortality for patients hospitalized with community-acquired pneumonia

Respir Res

Risk factors for community-acquired pneumonia in adults: a population-based case-control study

Eur Respir J

Angiotensin-converting enzyme inhibitor use and pneumonia risk in a general population

Eur Respir J

Original Article
International classification of diseases codes showed modest sensitivity for detecting community-acquired pneumonia