The Pediatric Infectious Disease Journal · 1999 · 70 citations · 11 references
MedicineInfectious DiseasesPediatricsInfant HealthNeonatologyPediatric Infectious DiseasesBacterial InfectionsCongenital InfectionsInfant Infectious DiseasesNeonatal InfectionsNeonatal Infectious DiseasesNeonatal SepsisBacterial Infections In NewbornsClinical Prediction RulesClinical Risk PredictionNeonatal Bacterial InfectionNeonatal Bacterial SepsisNeonatal Sepsis DiagnosisSerious Bacterial Infection
It is estimated that 5 million deaths occur in the neonatal period annually. Of these, ∼97% are in developing countries and >40% are believed to be caused by infection.1 In developing countries much of the effort to control neonatal mortality has concentrated on reducing the risk of perinatal infection by improving the care of pregnant mothers. For example referral of high risk pregnancies, maternal immunization for tetanus, improved care of infants after delivery with emphasis on umbilical cord care and the early introduction of breast-feeding. The treatment of infants with established infection remains unsatisfactory. Many cases of neonatal infection never reach treatment facilities, and the case-fatality rate for those that do ranges from 13 to 69%.2 One approach that has been used in dealing with other pediatric problems such as diarrhea and pneumonia has been to train clinicians, nurses and local health workers (e.g. traditional birth attendants or community health workers) to recognize clinical signs of illness and to refer children for treatment. In developing countries treatment decisions are based largely on the history and clinical examination because of the limited availability of laboratory facilities. More accurate methods of diagnosis could help improve decisions about the need for additional diagnostic tests or therapy in developing country settings. Most previous studies of the accuracy of the clinical examination in detecting serious bacterial illness in infants younger than 3 months of age have been conducted in developed countries. Studies in the US, United Kingdom and Australia3-5 have given conflicting results about the accuracy of observational scales in detecting serious bacterial illness. The small numbers of patients with serious bacterial illness who can be enrolled in developed countries have limited previous studies. This has made it difficult to evaluate the importance of large numbers of clinical variables and to use sophisticated statistical approaches to evaluate the relationship between clinical findings and diagnoses. In addition previous studies of serious bacterial illness have focused simply on the presence or absence of infection, an outcome that is less clinically relevant than one that also enables clinicians to assess the severity of illness. Finally developed country studies have placed greater reliance on the value of laboratory examinations and less emphasis on the clinical examination.3, 6 In developing countries laboratory facilities to perform tests such as the blood count and chest radiograph are often unavailable, and clinical decisions must be made without them. The purpose of this study was to develop a clinical prediction instrument that would identify infants at risk of serious bacterial illness on the basis of clinical examination findings. We hoped that the study would lead to a relatively simple method that could be used by nonphysicians and physicians to detect children at increased risk of adverse outcomes of serious bacterial illness. METHODS The design of the study has been described in detail previously.7 Briefly it was conducted at hospitals or outpatient clinics in Ethiopia, The Gambia, Papua New Guinea and The Philippines that see large numbers of sick infants. At each site infants younger than 91 days of age seen consecutively for acute care with chief complaints indicating possible infection were eligible to be included in the study. Entry criteria were intended to include infants with a wide spectrum of illness severity and to ensure that virtually all infants with serious infection would be included. Exclusion criteria are described elsewhere in this supplement.7 All infants underwent a standardized history and physical examination to assess the presence or absence and the degree of severity of signs and symptoms believed to be associated with bacterial disease (Table 1).8-10 Candidate predictors considered included demographic variables, historical variables, vital signs and physical examination findings. A pediatrician conducted the examination at three of the sites. Pediatric nurse practitioners performed the examination at the site in Papua New Guinea.TABLE 1: Clinical history and examination findings evaluated in model development Infants with prespecified symptoms associated with possible bacterial infection underwent a laboratory evaluation that included blood culture, white blood cell count and chest radiograph. Specific criteria were also used to identify infants for lumbar puncture. All infants enrolled in the study underwent pulse oximetry. Urine cultures were not systematically obtained in all sites because it was not feasible to collect suprapubic specimens or to catheterize infants in all the settings. A random sample of patients not meeting criteria for laboratory testing underwent chest radiograph, blood culture and blood count to assess the frequency of bacterial infection among infants not meeting criteria for laboratory evaluation. Decisions regarding treatment were made on a clinical basis. Bacteremia was defined as the growth of a known pathogen in cultures of blood or cerebrospinal fluid (CSF). Meningitis was defined as a positive CSF culture or >10 polymorphonuclear cells in a nonbloody CSF specimen. Patients with grossly blood-stained lumbar punctures were classified as not having meningitis unless the CSF culture was positive. The criteria used to classify organisms as contaminants are described elsewhere.7 Oxygen saturation was measured after the clinical examination with a Nellcor N-200 (Nellcor Inc., Hayward, CA) pulse oximeter as previously described. Oxygen saturation measurements from sites above sea level were adjusted to correspond to readings at sea level. Pneumonia was diagnosed on the basis of chest radiographs interpreted without clinical information by a panel of three radiologists.7, 11 Study infants were defined as having pneumonia if all the radiologists who read the radiograph considered it to show probable or definite pneumonia. Infants with uninterpretable films or with missing radiographic information were classified as having a normal radiograph. Study outcome measure. The primary study outcome measure was an ordinal scale that summarized the presence or absence of disease, as well as its severity. This measure was selected because it would allow infants to be included in the clinical prediction rule who might have negative cultures (e.g. those with low oxygen saturation and positive radiographs), but who still might benefit from treatment because they were at increased risk of bacterial infection. Data about oxygen saturation, chest radiography, blood culture, CSF culture and cell count and mortality were combined to produce a measure that grouped diagnosis in increasing severity. Diagnoses were ranked in a hierarchic fashion based on their association with the most severe outcome, death. Diagnoses not strongly related to death were examined with respect to their association with the presence of a positive blood culture or CSF result. Outcome variables with similar associations with death or bacterial infection were included in the same category. This approach produced a ranking of severity with four levels of disease severity: (1) no abnormality; (2) mild hypoxemia (90% ≧ SaO2 < 95%) or radiologic pneumonia; (3) severe hypoxemia (SaO2 < 90%) or bacteremia or meningitis; and (4) death. As will be discussed later the main study analysis considered only a three level ordinal scale that did not include death as an explicit outcome category. From a statistical standpoint an ordinal ranking assumes that the variables can be ordered in terms of clinical importance or disease severity without assuming interval spacing between the variables. Such a procedure also makes more efficient use of data for statistical analyses. Children who died >7 days after discharge from the hospital were not counted as a death for the purposes of this study. Deriving the clinical prediction rule. The goal of the statistical analysis was to select a limited set of clinical predictors that could reasonably be used in clinical settings and that could predict infants at risk of serious bacterial illness or death. The analysis was designed to reduce the total number of variables in the model by grouping related subsets of clinical findings and summarizing each group of findings with a single score. We sought to avoid selecting variables to include in the model on the basis of their individual statistical associations with specific diagnoses because the multiple tests involved in this approach would exaggerate the strength of the relationship between clinical findings and the diagnoses. The statistical methods of the analysis are described in detail elsewhere.12 The derivation of the clinical prediction rule took place in three steps. Step 1. Clustering variables. Statistical variable clustering methods were used to place the 51 individual clinical signs and symptoms into groups or "clusters" so that the correlation of clinical findings within clusters was high and the between cluster correlation was low. The clustering procedure took place before the evaluation of the association between the clusters and the diagnoses. After the clustering procedure clinicians on the study team were asked to identify and rank the relative severity of the findings within the clusters. After this exercise a check was made to identify major disagreements between physicians' scoring and the outcome patterns. The only clustering that was rescored after examining patterns of the association with the ordinal scale was the auscultation cluster. Here a decision was made to use as the only variable the presence or absence of crepitations. Clinicians' ranking and scoring decisions were carried out before any analysis relating individual clinical signs to the diagnoses. We selected a scoring method in which the cluster score was the score of the most severely abnormal finding in the cluster. Measurements of vital signs (age, temperature, respiratory rate, weight) were not included in the clusters. The weight-for-age Z score was used as the measure of infant size. We made this decision because other anthropomorphic variables of size in this age group are highly correlated with weight-for-age. Once each cluster of clinical signs was scored, we entered the cluster scores into a logistic regression model. The clustering of signs thus resulted in significant reduction in the number of variables in the model so that fewer regression coefficients needed estimation. Step 2. Estimating the overall accuracy of the clinical findings. The association of the clusters of clinical variables, vital signs measurements and age with the diagnostic outcome was assessed by fitting a logistic regression model for an ordinal response variable (continuation ratio model) using penalized maximum likelihood estimation. Penalization is a method of discounting fitted regression coefficients to avoid overfitting. All two way interactions between predictors and the category of the ordinal response being were included in the model so that predictors could be included that might be associated with one level of the scale but not with For example the presence of might be more highly associated with chest radiographic of pneumonia but less strongly associated with a positive blood variables such as and respiratory rate were entered in the model the use of to not the model to the of a or association of variables with the ordinal methods were used to assess the of the model after for overfitting. For the the main emphasis was on developing a clinical prediction instrument that infants to a diagnostic category on the basis of chest radiography, blood and CSF cultures and oxygen saturation without to We were in this model because of the deaths in the of the study took place within of This model might identify children more to benefit from Infants who died were to level outcomes to the presence or absence of the other diagnoses. The of this was that about of deaths were classified in abnormal diagnostic category. The of the three level model was with that of the four level model using logistic regression Step the model for clinical of the model was by it as a which with fewer variables could be We used regression and to predict the of the model and variables that of the in the The of the procedure was to the correlation between the model and the model above with a in the measure of the or individual were so that the variable was the one that resulted in the in the correlation between the model and the model outcomes were not used as the variable in this model procedure in that variable in this way has a low of selecting the model because of problems of multiple and levels for of model We summarized the diagnostic accuracy of the model as its to between infants without any diagnostic and infants in the of adverse We the the for each level in the ordinal We used the overall likelihood ratio to the of information in the We have not for individual variables because of model methods were used to the of the model to infants with no from those in the or severe The procedure also the accuracy for caused by the of the model in the same set of data used to develop A total of infants less than 91 days of age were in of the criteria for in the study and underwent a physical examination and pulse oximetry. the demographic and the clinical of the study infants at of infants were enrolled in Ethiopia, The and the More infants than any of the other sites were enrolled in Papua New As described previously infants in Papua New Guinea were less to be or to have also the relationship between the presence of clinical findings and diagnostic For example the presence of a history of increased the of any hypoxemia or pneumonia; severe hypoxemia bacteremia or from to It increased the of a more serious bacteremia or from to of infants included in the study and the relationship of clinical findings at to diagnostic outcomes Of the infants in the study sample the prespecified criteria for laboratory evaluation and underwent blood culture and chest the criteria for a lumbar puncture. The infants with blood in the CSF were from the analysis of abnormal CSF The infants no clinical findings evaluation. The infants were classified into the diagnostic no abnormality; pneumonia or mild meningitis or severe and all deaths took place in the children were and were known to have died as Of children in the group not meeting criteria for laboratory for a evaluation with mild meningitis or severe of infants with such diagnoses. the frequency of serious infection in infants not eligible for laboratory infants were selected from those not eligible for laboratory evaluation to have blood and cultures The number of infants selected was less than because most in this group to blood the infants in laboratory tests were positive blood All the positive cultures in infants from The who with A infant a negative evaluation but the after with additional 6 infants oxygen saturation, one a on chest radiography, and the other two infants or chest radiographic findings. of the infants in the group We did not for the possible produced by the study design because the would be and because the random sample to have the frequency of adverse in the group of children not meeting criteria for laboratory evaluation. of the clinical prediction rule. For the the main emphasis was on the development of a clinical prediction instrument in which infants were to a diagnostic category without they The model three than because infants who died were to levels of severity to the presence or absence of other diagnoses. Infants who died were as no of any other serious or were missing data about laboratory and were to the no pneumonia or mild and meningitis or severe For infants into the no or hypoxemia virtually all were of having a serious bacterial illness by the clinicians for the In of it was not possible to a chest radiograph or a blood infants who would have the criteria to be included in a level have been The overall of this is to lead to a of the accuracy of the model. The variable clustering procedure to an of the 51 clinical history and examination findings into clusters of clusters were included in a model with the respiratory rate, and weight-for-age Z score to predict the diagnostic The model for infants into three levels death abnormality; or severe a of for infants with no from those with any and an of for infants with no from those with more serious the accuracy of the model we the number of infants with any relative to the number of infants without an at each of disease also the likelihood at a number of 1: the number of infants with no with those with any diagnostic at each level of the number of infants with no severe with those with a severe diagnostic at each level of were estimated by the the data were into and each were to the for the The ratio of for abnormal normal was and were with the use of the regression The and to the of the to the of the three level model was with one in which death was as a diagnostic category (Table The accuracy of the four level model was in detecting infants at of illness but the between the was A model with only vital respiratory rate and of the total in This model an of for any diagnostic We also the accuracy of the in the four study sites and in accuracy of three and four level logistic regression in detecting infants with outcome The for all infants with no outcome abnormality; infants with any outcome abnormality; infants with severe or and death. All are logistic of the model. the they would be difficult to use clinically because they would clinicians to assess a large number of clinical predictors and to cluster We the three diagnostic category model to its use in clinical settings by a group of findings that could be used as a for treatment by clinicians in small hospitals and by health workers in and outpatient regression was used to those groups of clinical findings that to in given by the model. This procedure was for groups of findings as well as for individual clinical findings. the procedure we signs that were to measure in infants (e.g. those to be and and those with low and The procedure was conducted for the model any and for the model a more serious This approach to two clinical prediction with similar to predict the presence of any A model that used clusters of clinical signs included three vital signs respiratory rate, the age and clusters of clinical respiratory of to and This model an of in the of the model and an of in any A model with individual clinical findings was by selecting a model from a model all the individual clinical findings. In this model was used for each clinical finding the findings could be used by as or The approach resulted in a model with specific clinical findings to history of definite chest to with history of in as well as respiratory rate, and for The model an to children with any in the ordinal scale of and an accuracy of the model. This was to a model to predict more serious In its the model a serious an of of the model in clinical settings. We developed a of the for use in clinical settings and the individual clinical findings were so that the value of each finding included was to its ratio in the model. For predictors we for ranges that with respiratory rate was associated with a of an on the the for respiratory rate has multiple all of the clinical findings are as each of the clinical findings the number of in the if the finding is and no if the finding is For example a infant with a of respiratory rate of weight-for-age Z score of and would a total of and a of of having at mild hypoxemia or pneumonia and a of of having severe bacteremia or of clinical prediction model to predict any of We also the accuracy of the 3 diagnostic category model with that of the clinical signs in the for the of the sick signs or to presence of severe chest presence of a from the umbilical and less than normal In detecting infants younger than days of age with any the sick criteria an of with the 3 level model of the study was the for the of infants months of age has been to include to well to the to and to as of possible serious infection. improve the of the This study that clinical examination is of value in infants at risk of serious bacterial illness and in the severity of of the most information from the vital signs and size. A limited number of clinical findings additional diagnostic The clinical findings can be combined into a instrument that is to be of use the of the it the information by using vital signs on a basis. The has the use of simple clinical signs in the of sick infants. have been studies of the accuracy of the The clinical findings included in the sick infant were similar to those in this study. The greater accuracy of the findings be to two of the the more use of an vital and more accurate for a finding is The clinical prediction instrument temperature, anthropomorphic measurements and respiratory rate as variables and thus the of information that variables are other previous studies of the value of clinical findings in detecting and the results that evaluation of the clinical can lead to more accurate information about which clinical findings to use and to them. Studies from developed countries in their of the value of the clinical examination in detecting serious bacterial illness in infants. A study by in a that an scale to detect serious bacterial was of limited diagnostic value in infants younger than months of age with bacterial illness. In a study by and that the presence of any of clinical findings chest a history of of a of and a of in illness or the need for hospital in children younger than 6 months of the study is not with the most of the same variables, vital signs which were not considered by were as findings with previous studies in developed countries for we enrolled a large number of of severe diagnostic The large number of the use of more sophisticated statistical than has previously been This to more the relationship of the clinical findings to diagnoses. The approach to model development we used also to of clinical findings to such diagnoses. variables to a model the relationship is results in of information and variables with to the such as vital information and be for the greater of the clinical prediction rule. study from a number of of the data were from Papua New the site with the of severely We to data from this site with other data because was no a that the clinical would have a relationship to serious bacterial illness than those was the of specimens for A to identify have the number of infants in bacterial illness was This would have the of the clinical findings. Finally the model has not been in a sample of methods that the rule is to perform well in other is needed in the we designed the study to assess the importance of by blood cultures and chest radiographs on infants who did not the criteria for were fewer infants than on tests were The of infants with positive laboratory results in this sample that they have more significant disease than the group from which they were as the of infants were more to will need to develop approaches to or patients so that the importance of can be the approach to model development that we and the decisions we made regarding variable are to have the diagnostic accuracy of the model. This study of the diagnostic value of clinical findings that can be on The approach we developed an basis for developing and thus the methods that the to develop diagnostic In addition the study that an approach that the of diagnoses to disease severity than into groups is feasible and a more approach for studies. Finally the model so that the of a diagnosis can be simply from a will allow clinicians to for multiple clinical findings at and the of each of diagnoses for any given in a developing country laboratory testing is not the use of the clinical prediction model could improve the of clinicians to identify sick infants. the in accuracy the the that children with serious bacterial illness will be In developing country settings the of could be by of infants classified as low In developed countries information from the clinical prediction instrument could be combined with data from the clinical laboratory to diagnostic of the model in high and low as well as studies to to the and of the prediction instrument in clinical settings and by levels of health be conducted in the We the and workers at each of the study sites who so much to this but who are to The of and the of was to the of this In and made significant to this Finally we the and of the more than of sick infants enrolled in the study sites who the of this study at a the were
11
Observation Scales to Identify Serious Illness in Febrile Children
Paul L. McCarthy, Michael R. Sharpe, Sydney Z. Spiesel et al. · PEDIATRICS · 1982 · 362 citations
Nursing, Child Health, Pediatrics +11
The Global Impact of Neonatal Infection
Barbara J. Stoll · Clinics in Perinatology · 1997 · 358 citations