This post is part of The Pump Handle’s new “Public Health Classics” series exploring some of the classic studies and reports that have shaped the field of public health. Links to past posts in the series are available here. If you have a favorite Public Health Classic to recommend, let us know in the comments. And if you’re interested in contributing a post to the series, email us at thepumphandle@gmail.com (send us a link to the report or study along with a sentence or two about what you find most interesting or important about it).

By Sara Gorman

In the late 1970s and early 1980s, acute respiratory infections were a serious, though understudied, cause of mortality among young children. During this time period, it was estimated that the annual incidence of pneumonia among children under 5 was 2-4 per 100 in developed countries and 10-20 per 100 in developing countries. Between 1970 and 1979 at the Goroka Base Hospital in Papua New Guinea, pneumonia was responsible for 31% of 1040 child deaths in the hospital. Death due to pneumonia in children is usually caused by bacterial infection and is therefore eminently preventable by administering antibiotics. In developing countries, however, antibiotics may be in short supply, and there are risks to overtreatment of children with antibiotics.

Despite the recognition that pneumonia was causing many childhood deaths, there were no good criteria for establishing the need for antibiotic treatment or hospital admission among children presenting with symptoms of acute respiratory infections, most prominently cough. The problem was especially dire in developing countries where resources were limited. In response to this lack of protocol, Frank Shann and colleagues conducted a prospective study of 350 children in Papua New Guinea to determine the most reliable signs of need for antibiotics and intensive care in the hospital. The study developed protocols that have since become the basis of controlling and treating acute respiratory infections in children.

In this landmark 1984 study, Shann et al. examined 350 children: 200 pediatric outpatients presenting with cough, 100 age-matched controls without cough, and 50 children admitted to Goroka Hospital with pneumonia. The authors acknowledged that the gold standard for making a diagnosis of pneumonia is the finding of consolidation (liquid-filled tissue) on chest X-ray, X-rays were not readily available in their clinical setting. The next best sign for diagnosis at the time was the detection of crepitations (crackling sounds) while listening to the patient breathing with a stethoscope. However, as the authors point out, detecting crepitations requires skills primary health care workers may not have developed.

Shann et al. were looking for a diagnostic measure that did not rely on sophisticated equipment or subjective diagnostic skills that require extensive training, and they found one. The study showed that a respiratory rate greater than 50—obtained by counting the number of breaths per minute—served as a good proxy measure for crepitations and had optimal sensitivity and selectivity in predicting which patients with cough require antibiotic therapy.  Although not perfect, the use of a respiratory rate of >50 breaths per minute limited both the number of children treated unnecessarily with antibiotics (false positives) and the number of children requiring antibiotics who would be denied them (false negatives). Interestingly, fever was a poor indicator of creptitations in the study and therefore rejected as a marker of the need for antibiotic intervention. On the other hand, a mother’s report that her child was breathing rapidly was almost as accurate as respiratory rate in determining the need for antibiotics.

In addition to determining the necessity of antibiotic treatment, determining the need for hospital admission for acute respiratory infections, especially in resource-poor settings, was another major point of concern of the study. Shann and colleagues discovered that chest indrawing (retraction of the sternum with each breath) was a good predictor of severe illness and the need for admission to the hospital for intensive care. Given that the vast majority of outpatients with a cough recovered and did not require hospital admission, while only 2% of the outpatient sample presented with chest indrawing, the use of indrawing as a measure of the necessity of hospital admission succeeded in reducing unnecessary hospitalizations.

At the end of the study, Shann and colleagues made a plea for further research to determine which clinical signs best predict the need for antibiotic treatment and hospitalization. This plea was in fact largely answered by subsequent studies. Even in the face of advances in protocols for treating acute respiratory infections as well as some disagreement surrounding the usefulness of Shann et al’s criteria for antibiotic treatment and hospital admission, by 1994 the protocol had been widely adopted with some important modification (namely different threshold breathing rates for children of different ages). Intervention studies conducted in Bangladesh, India, Indonesia, Nepal, Pakistan, the Philippines, and Tanzania confirmed the results of Shann’s study and gave credibility to its protocols. Research in developed countries, in which chest X-rays could be used to validate clinical criteria, resulted in some modifications to the Shann et al criteria, such as lowering the respiratory rate cut-off for antibiotic therapy to 40 breaths per minute for children 1 to 4 years old and raising it to 60 breaths for minute for infants under 2 months of age.

Significantly, Shann’s protocols have been shown to be reliable indicators of the best course of treatment for children with respiratory infections in the poorest and most remote rural areas. Thus Shann et al’s simple study became the foundation for a global approach to treating acute respiratory infections in children, to working toward reduction in childhood deaths from pneumonia, and to the judicious use of antibiotics and hospital beds.

Sara Gorman is a PhD candidate at Harvard University. She has written extensively about HIV, TB, and women’s and children’s health for a variety of public health organizations, including Save a Mother and Boston Center for Refugee Health and Human Rights. She most recently worked in the policy division at the HIV Law Project.