Role of acute Illness observation scale (AIOS) in managing severe childhood pneumonia
http://www.100md.com
《美国医学杂志》
1 Civil Hospital Rohru, District Shimla, Himachal Pradesh, India
2 Government Medical College, Chandigarh, India
Objective. In the perspective of integrated management of childhood illness (IMCI) strategy and recent evidence favoring use of oral antibiotics in severe pneumonia, a generic illness severity index - Acute Illness Observation Scale (AIOS)- was prospectively validated in children with severe pneumonia in a civil hospital in remote hilly region. Methods. AIOS was used in quantifying overall severity of illness for eighty-nine consecutive children (age, 2-59 months) hospitalized with community-acquired severe pneumonia. A detailed clinimetric evaluation of scale was carried out and logistic regression analyses predicted the following outcomes: 1) mode of initial antimicrobial therapy (oral vs. parenteral); and 2) need for intravenous fluids at admission. Results. Majority of children (80.9%) with severe pneumonia scored abnormally (AIOS score >10) at initial evaluation. Children with abnormal AIOS scores (>10) had significantly greater severity of respiratory distress and higher incidence of radiological pneumonia. Outcome measures i.e. time to defervescence and length of hospital stay were also positively and significantly correlated with the scores. The six-item scale had good internal consistency (Cronbach's alpha 0.81); and its factor analysis yielded a single latent factor explaining 54% of variance in illness severity at admission. Furthermore, logistic regression analyses revealed an independent predictive ability of AIOS in aiding clinician to decide the mode of initial antimicrobial therapy (oral or parenteral), as well as need for intravenous fluids. Conclusion. Authors study indicates the clinimetric validity of AIOS in managing, Severe childhood pneumonia and suggests its role in further enriching IMCI strategy.
Keywords: Acute Illness observation scale (AIOS); Clinimetrics; Community-acquired pneumonia; Illness severity; Integrated management of childhood Illness (IMCI)
Gaining an objective understanding of well-being of a child with pneumonia is essential to optimize criteria for triage, early referral, hospitalization and initial therapeutic modalities in less-developed countries.[1],[2] This has been much aided by Integrated Management of Childhood Illness (IMCI) strategy that simplifies the classification of illness severity for major acute childhood illnesses including pneumonia.[3] However, IMCI strategy will be more effective in managing pneumonia when supplemented by an illness-severity scoring system delivered in the context of primary care setting that can quantify quickly the severity of illness at all stages from onset to recovery. This need has been augmented by the recent evidence favoring oral antibiotic in treatment of severe community acquired pneumonia.[4] An objective and graded appraisal of "clinical appearance" easily ascertainable by primary care givers, can be instrumental in influencing subsequent management decisions.[5],[6],[7] In this regard, use of Acute Illness Observation Scale (AIOS) - a generic illness severity scale- represents a distinctive paradigm drawing on simple observations (based on toxic appearance) instead of complex symptomatology, aiming for wholeness rather than details and encompassing the entire not just the ends of sickness continuum.[8],[9] AIOS (a three-point scale for six ordinal variables and total score range of 6 to 30) is a validated clinical index of quantifying risk of serious bacterial infections (SBI) in children 36 months or younger presenting with febrile illnesses.[10] Incidence of SBI is less than 2-3% if a febrile child scores 10 or less; and more than 90% if AIOS score is 16 or above.[6],[10]
Therefore, the aim of our study is to validate AIOS in predicting illness severity, initial therapeutic modalities, and clinical outcomes of severe community acquired pneumonia, by use of simple observational variables available at the time of triage even before physical examination; and to compare the predictive performance of this model with those of specific respiratory signs in deciding the mode of antimicrobial therapy and need for intravenous fluids by primary care giver.
Material and Methods
Study participants and setting
Children aged 2 to 59 months with diagnosis of community-acquired severe pneumonia (based on WHO criteria) [1] were consecutively enrolled. Only those patients who received at least 72 hours of initial parenteral antimicrobials before switching over to oral therapy were included in parenteral antimicrobial group; while oral antimicrobial group included children who were put on oral chloramphenicol either right at admission or after they had received initial parenteral antimicrobials for not more than 12 hours. The children who received initial parenteral antimicrobials for at least 24-48 hours before switching over to oral antibiotics were excluded from the study, so as to make the outcome parameter of initial antibiotic therapy more clearly definable and undeniably related to severity of pneumonia. The following were other reasons for exclusion: children having duration of illness >2 weeks; or respiratory distress with predominant wheezing. Patients were recruited prospectively between May 1997 and June 1998 at a civil hospital (tehsil Rohroo, district Shimla, Himachal Pradesh, India).
Though all children with severe pneumonia were originally planned to undergo radiographic examination within 24 hours of admission, only 83.1% (74/89) children had their chest radiographs (CXRs) done due to logistic problems. A single radiologist, who was blind to the clinical diagnosis, evaluated the radiographs.[11] As far as treatment is concerned, oral antimicrobial was prescribed if any 2 of the following criteria were present: 1) preserved general status, based on clinicians' overall impression; 2) less severe respiratory distress i.e. absence of severe recessions, marked grunting and inability to drink; or 3) normal CXR. Children who failed to meet above-mentioned criteria were started on parenteral antibiotics.
Acute Illness Observation Scale (AIOS) [10] was meticulously recorded for each affected child in a reasonably quite state. The two clinical researchers themselves meticulously recorded the scores at presentation; and documented clinical course as well as resolution of symptoms and signs throughout the length of stay in hospital. Clinimetric evaluation of AIOS included item analysis, reliability measurement (internal consistency reliability) and validity (construct and predictive) testing. All P values reported in this study are 2-sided unless indicated otherwise.
Results
Our study sample consisted of 89 children, 55% of whom were male. They ranged in age from 2.4 to 59 months (mean, 19.7; median 12 months); and infants (2-12 months) (51.6%) being most frequently affected. All children presented with complaints of fever and cough while history of rapid or difficult breathing was obtained only in 57% of cases. The duration of most common presenting complaints and incidence of various signs of respiratory distress stratified by the type of antimicrobial treatment are listed in [Table - 1].
Due to logistic problems, radiographic examination of chest was carried out only in 83.1% (74/89) children, and 4 children were further excluded from analysis due to poor quality of their radiographs. Normal CXRs were present in 21.4% (15/70) and remaining 78.5% (55/70) had significant radiological abnormalities [Table - 1].
During their management, 31.5% (28) children received intravenous fluids because of severe respiratory distress (mainly inability to drink, drowsiness, or marked grunting) and/or dehydration. Parenteral antimicrobials (penicillin/ampicillin + chloramphenicol) were administered to 64% (57) patients, while remaining 36%(32) were treated with oral chloramphenicol mainly. Severe shortage of oxygen limited its use to only those 2 infants who had obvious cyanosis. The mean duration (± SD) of hospitalization was 6.4 (± 2.8) days.
Illness Severity Scale (AIOS) and its Clinimetrics
The composition and scoring pattern of AIOS scale with its clinical significance are presented in [Table - 2]. Seventy-two children (80.9%) with pneumonia scored abnormally (AIOS score >10) at initial evaluation. Mean as well as median scores for AIOS clearly signify the seriousness of all children enrolled in the study. The frequency of abnormal AIOS scores as well as mean total scores were not significantly different in children older than 3 years (10/15 [66.6%] had AIOS score >10; mean total score, 14.5 ± 5.5), in contrast to children aged 3 or less (62/74 [83.7%] had AIOS score >10; mean total score, 16.6 ± 5.4).
As far as the individual item analysis of AIOS is concerned, 61.8% and 58.4% of affected children scored normally for the variables "color" and "hydration" respectively (median score of 1 each). In contrast, majority of affected children showed worst outcome in "response to social overtures" variable (i.e.79.8% had median score of 5). In the score range of 1 to 5, median scores of the affected children clearly indicate floor and ceiling effects in these variables respectively.
Moreover, scales were assessed for their inter-item correlations, corrected item to total correlations, and overall Cronbach's a (only latter is shown in [Table - 2]). Cronbach's a for AIOS was 0.81(an alpha of 0.70 is the minimum desirable level) indicating the homogeneity of scale variables in assessing illness severity in our study sample.[12] Overall, the individual item analysis of AIOS revealed either similar or decreased values for a-if-item-deleted (range 0.77-0.81), indicating that each item added unique information to the total score.
Another key clinimetric feature of AIOS evaluated in this study was its construct validity. The total score on AIOS showed highly significant correlations (Spearman) with selected clinical characteristics at admission like grade of fever (P <. 01), respiratory rate (P <. 01), heart rate (P <. 01) and a significant trend for the variable of treatment duration prior to admission (P <. 1) (data are not shown). Factor analysis extracted only a single factor with an eigenvalue of 3.2 (minimum cut off being 1), [13] which accounted for 54% of variance in the whole model. All the 6 variables of the AIOS had higher factor loadings (0.6 or >) for a single latent component confirming thereby its unidimensional nature as far as appraisal of general severity of illness in pneumonia is concerned.
Relating children's scores against their radiologic findings to assess the concurrent validity, 40% (6/15) children with normal CXR findings had AIOS of 10 or less whereas only 14.5% (8/55) had normal scores in the group with abnormal CXR findings (c 2sub: 4.77; P=0.029); though their mean scores were not significantly different (15.33 ± 8.16 versus 16.47 ± 4.92; P=0.61). On the other hand, severity of respiratory distress (i.e. respiratory rate, grunting, severe recessions and inability to drink) was similar between children with normal and abnormal chest radiographs.
Respiratory distress- and treatment-related variables of affected children were also stratified by their illness severity scores at presentation. Children scoring abnormally on AIOS (>10) had significantly higher frequency of severe tachypnea (P <0.01), marked recessions (P < .05), and grunting (P < 0.01) while frequency of inability to drink reached statistical significance (P <0.05) only for children who scored 3 16 on AIOS [Table - 3]. Furthermore, frequency of administration of intravenous fluids and parenteral antibiotics was significantly higher in those children with severe pneumonia whose scores in both the scales were abnormal [Table - 3]. In addition, the Spearman correlations for clinical outcome measures (time to defervescence and length of hospital stay) indicate that higher the scores on AIOS, longer it took for tachypnea to decrease (P < .01) as well as subside (P < .01) and hospital stay was also prolonged (P < .01)(data not shown). Although not significant, scores also tended to positively correlate with time taken for fever to settle (P < .10).
We also used logistic regression analyses to examine relationship of these two dichotomous therapeutic measures (mode of initial antimicrobial therapy i.e., oral vs. parenteral; and need for intravenous fluids) with selected predictors [Table - 4]. These predictors were tachypnea, severe recessions, grunting, inability to drink and AIOS. In these analyses, the effects of predictors were evaluated one at a time (models 1 through 5) and in the group (model 6). All the predictors were entered in these analyses as binary variables except tachypnea. AIOS was dichotomized as ￡ 10 vs. >10 for the outcome of oral vs. parenteral antimicrobials; and <16 vs. 316 for the outcome of need for intravenous fluids.
[Table - 4] shows that illness severity scale (AIOS) as well as grunting and severe recessions significantly predicted the mode of initial antimicrobial therapy (oral vs parenteral) at admission and significance of AIOS scale (with cut- off >10) was maintained even when parameters like grunting and recessions were entered along in model 6. Similarly, when tachypnea, severe recessions, grunting and inability to drink were entered along with AIOS severity scale in model 6 for the outcome of need for intravenous fluids in children with severe pneumonia, AIOS (with cutoff score 3 16) remained statistically significant and independent predictor [Table - 4].
Discussion
This study had demonstrated high internal consistency as well as external validity of AIOS in children with severe community-acquired pneumonia. The compromised general status entailing various observation variables of AIOS had already shown to be a significant and independent predictor of serious illnesses including hypoxemia in respiratory tract infections in developing countries.[5],[7],[14] Importantly, all the three components of care envisaged in IMCI strategy can be upgraded by use of AIOS. First, the evidence based syndromic approach lays significant emphasis on evaluating severity of child's condition by primary care workers who usually misclassify symptoms with overlapping causes or for which a single diagnosis may not be appropriate using earlier vertical disease WHO algorithms.[15] AIOS seems to fulfill this role in a simple and objective manner. In a series of articles beginning in 1980,[8],[9],[10],[16],[17] McCarthy et al had already demonstrated the utility of AIOS in identifying those febrile children who have the most toxic illness and those who have serious illness (e.g. pneumonia, UTI, meningitis, severe gastroenteritis, a focal complication etc.). Studies criticizing AIOS were mainly restricted to babies below 8 weeks of age, [18] and those with occult bacteremia in nontoxic children.[19] AIOS offers an explicit, objective, and actionable criterion that could be easily implemented in real world practice. Second, the in-hospital curatives services also can be rationed by use of AIOS, which might safely increase the proportion of children with severe community, acquired pneumonia that can be treated as outpatients with oral antibiotics. Lastly, it also can boost the skills of mother to identify sickness of a child at home. In this regard, a randomized trial aimed at educating parents about the use of AIOS, [20] had demonstrated that its use results in more reliable parent judgment about well being of children during acute illnesses.
With respect to aims of the current study, some methodological issues need to be addressed. In the absence of in-hospital mortality as well as facility for pulse oximetry, the outcomes used for logistic regression analyses were somewhat soft and less objective. However, by clearly defining the initial antimicrobial therapy groups and indication for intravenous hydration, the outcome parameters are easily definable and ascertainable. Lack of microbiological support also prevented us from assessing the performance of AIOS in relation to the etiology of pneumonia. Other areas requiring further work include determining feasibility and reproducibility of AIOS by less qualified workers, elucidating its discriminative sensitivity in severe and non severe pneumonia, testing the responsiveness of scale to change during treatment and optimizing its diagnostic accuracy (vis a vis pulse oximetry or arterial blood gas analysis).
To conclude finally, the features of respiratory distress are no doubt, any clinician's first intentions in a child with pneumonia, yet discerning the overall sickness in an objective manner by a simple clinical index like AIOS with its six observation variables may behoove the treating clinician or primary care giver closer to the finality of the disease severity; and further optimize IMCI strategy for community management of childhood acute lower respiratory tract infections.
References
1.WHO Programme for the Control of Acute Respiratory Infections . Acute respiratory infections in children: case management in small hospitals in developing countries. Geneva: World Health Organization, 1990.
2.Nolan T, Angos P, Cunha A JLA, Muhe L, Qazi S, Simoes EAF et al. Quality of hospital care for seriously ill children in less-developed countries. Lancet 2001; 357 : 106-110.
3.Gove S. Integrated management of childhood illness by outpatient health workers: technical basis and overview. Bull World Health Org 1997; 75 (suppl 1): S7-S 24.
4.Addo-Yobo E, Chisaka N, Hassan M, Hibberd P, Lozano J M, Jeena P et al. Oral amoxicillin versus injectable penicillin for severe pneumonia in children aged 3 to 59 months: a randomized multicentre equivalency study. Lancet 2004; 364 : 1141-1148.
5.Bang AT, Bang RA, Reddy MH, Baitule SB, Deshmukh MD, Paul VK et al. Simple clinical criteria to identify sepsis or pneumonia in neonates in the community needing treatment or referral. Pediatr Infect Dis J 2005; 24 : 335-341.
6.Bonadio WA. The history and physical assessments of the febrile infant. Pediatr Clin North Am 1998; 45 : 65-77. [PUBMED]
7.Weber MW, Usen S, Palmer A, Jaffar S, Mulholland EK. Predictors of hypoxaemia in hospital admissions with acute lower respiratory tract infection in a developing country. Arch Dis Child 1997; 76 : 310-314.
8.McCarthy PL, Jekel JF, Stashwick CA, Spiesel SZ, Dolan TF Jr. History and observation variables in assessing febrile children . Pediatrics 1980; 65 : 1090-1095. [PUBMED]
9.McCarthy PL, Jekel JF, Stashwick CA, Spiesel SZ, Dolan TF, Sharpe MR et al. Further definition of history and observation variables in assessing febrile children. Pediatrics 1981; 67:687-93.
10.McCarthy PL, Sharpe MR, Spiesel SZ, Dolan TF, Forsyth BW, DeWitt TG et al. Observation scales to identify serious illness in febrile children. Pediatrics 1982; 70 : 802-809.
11.World Health Organization Pneumonia Vaccine Trial Investigators' Group. Standardization of interpretation of chest radiographs for the diagnosis of pneumonia in children . Geneva: World Health Organization, 2001.
12.Knapp TR. Coefficient alpha: conceptualizations and anomalies. Res Nurs Hlth 1991; 14:457-60.
13.Ferketich S and Muller M. Factor analysis revisited. Nurs Res 1990; 39 : 59-62.
14.Rajesh VT, Singhi S, Kataria S. Tachypnea is a good predictor of hypoxia in acutely ill children. Arch Dis Child 2000; 82 : 46-9. [PUBMED] [FULLTEXT]
15.Costello A. Is India ready for the integrated management of childhood illness strategy Indian Pediatr 1999; 36: 759-62. [PUBMED] [FULLTEXT]
16.McCarthy PL, Lembo RM, Baron MA, Fink HD, Cicchetti DV. Predictive value of abnormal physical examination findings in ill-appearing and well-appearing febrile children. Pediatrics 1985; 76 : 167-171. [PUBMED]
17.McCarthy PL, Lembo RM, Fink HD, Baron MA, Cicchetti DV. Observation, history, and physical examination in diagnosis of serious illnesses in febrile children = 24 months. J Pediatr 1987; 110 : 26-30. [PUBMED]
18.Baker MD, Avner JR, Bell LM. Failure of infant observation scales in detecting serious illness in febrile, 4-to 8-week- old infants. Pediatrics 1990; 85 : 1040-1043. [PUBMED]
19.Teach SJ, Fleisher GR, Occult Bacteremia Study Group. Efficacy of an observation scale in detecting bacteremia in febrile children three to thirty-six months of age, treated as outpatients. J Pediatr 1995; 126 : 877-881.
20.McCarthy PL, Sznajderman SD, Lustman-Findling K, Baron MA, Fink HD, Czarkowski N et al. Mothers' clinical judgment: a randomized trial of the Acute Illness Observation Scales. J Pediatr 1990; 116 : 200-206.(Bharti Bhavneet, Bharti Sahul, Verma Van)
2 Government Medical College, Chandigarh, India
Objective. In the perspective of integrated management of childhood illness (IMCI) strategy and recent evidence favoring use of oral antibiotics in severe pneumonia, a generic illness severity index - Acute Illness Observation Scale (AIOS)- was prospectively validated in children with severe pneumonia in a civil hospital in remote hilly region. Methods. AIOS was used in quantifying overall severity of illness for eighty-nine consecutive children (age, 2-59 months) hospitalized with community-acquired severe pneumonia. A detailed clinimetric evaluation of scale was carried out and logistic regression analyses predicted the following outcomes: 1) mode of initial antimicrobial therapy (oral vs. parenteral); and 2) need for intravenous fluids at admission. Results. Majority of children (80.9%) with severe pneumonia scored abnormally (AIOS score >10) at initial evaluation. Children with abnormal AIOS scores (>10) had significantly greater severity of respiratory distress and higher incidence of radiological pneumonia. Outcome measures i.e. time to defervescence and length of hospital stay were also positively and significantly correlated with the scores. The six-item scale had good internal consistency (Cronbach's alpha 0.81); and its factor analysis yielded a single latent factor explaining 54% of variance in illness severity at admission. Furthermore, logistic regression analyses revealed an independent predictive ability of AIOS in aiding clinician to decide the mode of initial antimicrobial therapy (oral or parenteral), as well as need for intravenous fluids. Conclusion. Authors study indicates the clinimetric validity of AIOS in managing, Severe childhood pneumonia and suggests its role in further enriching IMCI strategy.
Keywords: Acute Illness observation scale (AIOS); Clinimetrics; Community-acquired pneumonia; Illness severity; Integrated management of childhood Illness (IMCI)
Gaining an objective understanding of well-being of a child with pneumonia is essential to optimize criteria for triage, early referral, hospitalization and initial therapeutic modalities in less-developed countries.[1],[2] This has been much aided by Integrated Management of Childhood Illness (IMCI) strategy that simplifies the classification of illness severity for major acute childhood illnesses including pneumonia.[3] However, IMCI strategy will be more effective in managing pneumonia when supplemented by an illness-severity scoring system delivered in the context of primary care setting that can quantify quickly the severity of illness at all stages from onset to recovery. This need has been augmented by the recent evidence favoring oral antibiotic in treatment of severe community acquired pneumonia.[4] An objective and graded appraisal of "clinical appearance" easily ascertainable by primary care givers, can be instrumental in influencing subsequent management decisions.[5],[6],[7] In this regard, use of Acute Illness Observation Scale (AIOS) - a generic illness severity scale- represents a distinctive paradigm drawing on simple observations (based on toxic appearance) instead of complex symptomatology, aiming for wholeness rather than details and encompassing the entire not just the ends of sickness continuum.[8],[9] AIOS (a three-point scale for six ordinal variables and total score range of 6 to 30) is a validated clinical index of quantifying risk of serious bacterial infections (SBI) in children 36 months or younger presenting with febrile illnesses.[10] Incidence of SBI is less than 2-3% if a febrile child scores 10 or less; and more than 90% if AIOS score is 16 or above.[6],[10]
Therefore, the aim of our study is to validate AIOS in predicting illness severity, initial therapeutic modalities, and clinical outcomes of severe community acquired pneumonia, by use of simple observational variables available at the time of triage even before physical examination; and to compare the predictive performance of this model with those of specific respiratory signs in deciding the mode of antimicrobial therapy and need for intravenous fluids by primary care giver.
Material and Methods
Study participants and setting
Children aged 2 to 59 months with diagnosis of community-acquired severe pneumonia (based on WHO criteria) [1] were consecutively enrolled. Only those patients who received at least 72 hours of initial parenteral antimicrobials before switching over to oral therapy were included in parenteral antimicrobial group; while oral antimicrobial group included children who were put on oral chloramphenicol either right at admission or after they had received initial parenteral antimicrobials for not more than 12 hours. The children who received initial parenteral antimicrobials for at least 24-48 hours before switching over to oral antibiotics were excluded from the study, so as to make the outcome parameter of initial antibiotic therapy more clearly definable and undeniably related to severity of pneumonia. The following were other reasons for exclusion: children having duration of illness >2 weeks; or respiratory distress with predominant wheezing. Patients were recruited prospectively between May 1997 and June 1998 at a civil hospital (tehsil Rohroo, district Shimla, Himachal Pradesh, India).
Though all children with severe pneumonia were originally planned to undergo radiographic examination within 24 hours of admission, only 83.1% (74/89) children had their chest radiographs (CXRs) done due to logistic problems. A single radiologist, who was blind to the clinical diagnosis, evaluated the radiographs.[11] As far as treatment is concerned, oral antimicrobial was prescribed if any 2 of the following criteria were present: 1) preserved general status, based on clinicians' overall impression; 2) less severe respiratory distress i.e. absence of severe recessions, marked grunting and inability to drink; or 3) normal CXR. Children who failed to meet above-mentioned criteria were started on parenteral antibiotics.
Acute Illness Observation Scale (AIOS) [10] was meticulously recorded for each affected child in a reasonably quite state. The two clinical researchers themselves meticulously recorded the scores at presentation; and documented clinical course as well as resolution of symptoms and signs throughout the length of stay in hospital. Clinimetric evaluation of AIOS included item analysis, reliability measurement (internal consistency reliability) and validity (construct and predictive) testing. All P values reported in this study are 2-sided unless indicated otherwise.
Results
Our study sample consisted of 89 children, 55% of whom were male. They ranged in age from 2.4 to 59 months (mean, 19.7; median 12 months); and infants (2-12 months) (51.6%) being most frequently affected. All children presented with complaints of fever and cough while history of rapid or difficult breathing was obtained only in 57% of cases. The duration of most common presenting complaints and incidence of various signs of respiratory distress stratified by the type of antimicrobial treatment are listed in [Table - 1].
Due to logistic problems, radiographic examination of chest was carried out only in 83.1% (74/89) children, and 4 children were further excluded from analysis due to poor quality of their radiographs. Normal CXRs were present in 21.4% (15/70) and remaining 78.5% (55/70) had significant radiological abnormalities [Table - 1].
During their management, 31.5% (28) children received intravenous fluids because of severe respiratory distress (mainly inability to drink, drowsiness, or marked grunting) and/or dehydration. Parenteral antimicrobials (penicillin/ampicillin + chloramphenicol) were administered to 64% (57) patients, while remaining 36%(32) were treated with oral chloramphenicol mainly. Severe shortage of oxygen limited its use to only those 2 infants who had obvious cyanosis. The mean duration (± SD) of hospitalization was 6.4 (± 2.8) days.
Illness Severity Scale (AIOS) and its Clinimetrics
The composition and scoring pattern of AIOS scale with its clinical significance are presented in [Table - 2]. Seventy-two children (80.9%) with pneumonia scored abnormally (AIOS score >10) at initial evaluation. Mean as well as median scores for AIOS clearly signify the seriousness of all children enrolled in the study. The frequency of abnormal AIOS scores as well as mean total scores were not significantly different in children older than 3 years (10/15 [66.6%] had AIOS score >10; mean total score, 14.5 ± 5.5), in contrast to children aged 3 or less (62/74 [83.7%] had AIOS score >10; mean total score, 16.6 ± 5.4).
As far as the individual item analysis of AIOS is concerned, 61.8% and 58.4% of affected children scored normally for the variables "color" and "hydration" respectively (median score of 1 each). In contrast, majority of affected children showed worst outcome in "response to social overtures" variable (i.e.79.8% had median score of 5). In the score range of 1 to 5, median scores of the affected children clearly indicate floor and ceiling effects in these variables respectively.
Moreover, scales were assessed for their inter-item correlations, corrected item to total correlations, and overall Cronbach's a (only latter is shown in [Table - 2]). Cronbach's a for AIOS was 0.81(an alpha of 0.70 is the minimum desirable level) indicating the homogeneity of scale variables in assessing illness severity in our study sample.[12] Overall, the individual item analysis of AIOS revealed either similar or decreased values for a-if-item-deleted (range 0.77-0.81), indicating that each item added unique information to the total score.
Another key clinimetric feature of AIOS evaluated in this study was its construct validity. The total score on AIOS showed highly significant correlations (Spearman) with selected clinical characteristics at admission like grade of fever (P <. 01), respiratory rate (P <. 01), heart rate (P <. 01) and a significant trend for the variable of treatment duration prior to admission (P <. 1) (data are not shown). Factor analysis extracted only a single factor with an eigenvalue of 3.2 (minimum cut off being 1), [13] which accounted for 54% of variance in the whole model. All the 6 variables of the AIOS had higher factor loadings (0.6 or >) for a single latent component confirming thereby its unidimensional nature as far as appraisal of general severity of illness in pneumonia is concerned.
Relating children's scores against their radiologic findings to assess the concurrent validity, 40% (6/15) children with normal CXR findings had AIOS of 10 or less whereas only 14.5% (8/55) had normal scores in the group with abnormal CXR findings (c 2sub: 4.77; P=0.029); though their mean scores were not significantly different (15.33 ± 8.16 versus 16.47 ± 4.92; P=0.61). On the other hand, severity of respiratory distress (i.e. respiratory rate, grunting, severe recessions and inability to drink) was similar between children with normal and abnormal chest radiographs.
Respiratory distress- and treatment-related variables of affected children were also stratified by their illness severity scores at presentation. Children scoring abnormally on AIOS (>10) had significantly higher frequency of severe tachypnea (P <0.01), marked recessions (P < .05), and grunting (P < 0.01) while frequency of inability to drink reached statistical significance (P <0.05) only for children who scored 3 16 on AIOS [Table - 3]. Furthermore, frequency of administration of intravenous fluids and parenteral antibiotics was significantly higher in those children with severe pneumonia whose scores in both the scales were abnormal [Table - 3]. In addition, the Spearman correlations for clinical outcome measures (time to defervescence and length of hospital stay) indicate that higher the scores on AIOS, longer it took for tachypnea to decrease (P < .01) as well as subside (P < .01) and hospital stay was also prolonged (P < .01)(data not shown). Although not significant, scores also tended to positively correlate with time taken for fever to settle (P < .10).
We also used logistic regression analyses to examine relationship of these two dichotomous therapeutic measures (mode of initial antimicrobial therapy i.e., oral vs. parenteral; and need for intravenous fluids) with selected predictors [Table - 4]. These predictors were tachypnea, severe recessions, grunting, inability to drink and AIOS. In these analyses, the effects of predictors were evaluated one at a time (models 1 through 5) and in the group (model 6). All the predictors were entered in these analyses as binary variables except tachypnea. AIOS was dichotomized as ￡ 10 vs. >10 for the outcome of oral vs. parenteral antimicrobials; and <16 vs. 316 for the outcome of need for intravenous fluids.
[Table - 4] shows that illness severity scale (AIOS) as well as grunting and severe recessions significantly predicted the mode of initial antimicrobial therapy (oral vs parenteral) at admission and significance of AIOS scale (with cut- off >10) was maintained even when parameters like grunting and recessions were entered along in model 6. Similarly, when tachypnea, severe recessions, grunting and inability to drink were entered along with AIOS severity scale in model 6 for the outcome of need for intravenous fluids in children with severe pneumonia, AIOS (with cutoff score 3 16) remained statistically significant and independent predictor [Table - 4].
Discussion
This study had demonstrated high internal consistency as well as external validity of AIOS in children with severe community-acquired pneumonia. The compromised general status entailing various observation variables of AIOS had already shown to be a significant and independent predictor of serious illnesses including hypoxemia in respiratory tract infections in developing countries.[5],[7],[14] Importantly, all the three components of care envisaged in IMCI strategy can be upgraded by use of AIOS. First, the evidence based syndromic approach lays significant emphasis on evaluating severity of child's condition by primary care workers who usually misclassify symptoms with overlapping causes or for which a single diagnosis may not be appropriate using earlier vertical disease WHO algorithms.[15] AIOS seems to fulfill this role in a simple and objective manner. In a series of articles beginning in 1980,[8],[9],[10],[16],[17] McCarthy et al had already demonstrated the utility of AIOS in identifying those febrile children who have the most toxic illness and those who have serious illness (e.g. pneumonia, UTI, meningitis, severe gastroenteritis, a focal complication etc.). Studies criticizing AIOS were mainly restricted to babies below 8 weeks of age, [18] and those with occult bacteremia in nontoxic children.[19] AIOS offers an explicit, objective, and actionable criterion that could be easily implemented in real world practice. Second, the in-hospital curatives services also can be rationed by use of AIOS, which might safely increase the proportion of children with severe community, acquired pneumonia that can be treated as outpatients with oral antibiotics. Lastly, it also can boost the skills of mother to identify sickness of a child at home. In this regard, a randomized trial aimed at educating parents about the use of AIOS, [20] had demonstrated that its use results in more reliable parent judgment about well being of children during acute illnesses.
With respect to aims of the current study, some methodological issues need to be addressed. In the absence of in-hospital mortality as well as facility for pulse oximetry, the outcomes used for logistic regression analyses were somewhat soft and less objective. However, by clearly defining the initial antimicrobial therapy groups and indication for intravenous hydration, the outcome parameters are easily definable and ascertainable. Lack of microbiological support also prevented us from assessing the performance of AIOS in relation to the etiology of pneumonia. Other areas requiring further work include determining feasibility and reproducibility of AIOS by less qualified workers, elucidating its discriminative sensitivity in severe and non severe pneumonia, testing the responsiveness of scale to change during treatment and optimizing its diagnostic accuracy (vis a vis pulse oximetry or arterial blood gas analysis).
To conclude finally, the features of respiratory distress are no doubt, any clinician's first intentions in a child with pneumonia, yet discerning the overall sickness in an objective manner by a simple clinical index like AIOS with its six observation variables may behoove the treating clinician or primary care giver closer to the finality of the disease severity; and further optimize IMCI strategy for community management of childhood acute lower respiratory tract infections.
References
1.WHO Programme for the Control of Acute Respiratory Infections . Acute respiratory infections in children: case management in small hospitals in developing countries. Geneva: World Health Organization, 1990.
2.Nolan T, Angos P, Cunha A JLA, Muhe L, Qazi S, Simoes EAF et al. Quality of hospital care for seriously ill children in less-developed countries. Lancet 2001; 357 : 106-110.
3.Gove S. Integrated management of childhood illness by outpatient health workers: technical basis and overview. Bull World Health Org 1997; 75 (suppl 1): S7-S 24.
4.Addo-Yobo E, Chisaka N, Hassan M, Hibberd P, Lozano J M, Jeena P et al. Oral amoxicillin versus injectable penicillin for severe pneumonia in children aged 3 to 59 months: a randomized multicentre equivalency study. Lancet 2004; 364 : 1141-1148.
5.Bang AT, Bang RA, Reddy MH, Baitule SB, Deshmukh MD, Paul VK et al. Simple clinical criteria to identify sepsis or pneumonia in neonates in the community needing treatment or referral. Pediatr Infect Dis J 2005; 24 : 335-341.
6.Bonadio WA. The history and physical assessments of the febrile infant. Pediatr Clin North Am 1998; 45 : 65-77. [PUBMED]
7.Weber MW, Usen S, Palmer A, Jaffar S, Mulholland EK. Predictors of hypoxaemia in hospital admissions with acute lower respiratory tract infection in a developing country. Arch Dis Child 1997; 76 : 310-314.
8.McCarthy PL, Jekel JF, Stashwick CA, Spiesel SZ, Dolan TF Jr. History and observation variables in assessing febrile children . Pediatrics 1980; 65 : 1090-1095. [PUBMED]
9.McCarthy PL, Jekel JF, Stashwick CA, Spiesel SZ, Dolan TF, Sharpe MR et al. Further definition of history and observation variables in assessing febrile children. Pediatrics 1981; 67:687-93.
10.McCarthy PL, Sharpe MR, Spiesel SZ, Dolan TF, Forsyth BW, DeWitt TG et al. Observation scales to identify serious illness in febrile children. Pediatrics 1982; 70 : 802-809.
11.World Health Organization Pneumonia Vaccine Trial Investigators' Group. Standardization of interpretation of chest radiographs for the diagnosis of pneumonia in children . Geneva: World Health Organization, 2001.
12.Knapp TR. Coefficient alpha: conceptualizations and anomalies. Res Nurs Hlth 1991; 14:457-60.
13.Ferketich S and Muller M. Factor analysis revisited. Nurs Res 1990; 39 : 59-62.
14.Rajesh VT, Singhi S, Kataria S. Tachypnea is a good predictor of hypoxia in acutely ill children. Arch Dis Child 2000; 82 : 46-9. [PUBMED] [FULLTEXT]
15.Costello A. Is India ready for the integrated management of childhood illness strategy Indian Pediatr 1999; 36: 759-62. [PUBMED] [FULLTEXT]
16.McCarthy PL, Lembo RM, Baron MA, Fink HD, Cicchetti DV. Predictive value of abnormal physical examination findings in ill-appearing and well-appearing febrile children. Pediatrics 1985; 76 : 167-171. [PUBMED]
17.McCarthy PL, Lembo RM, Fink HD, Baron MA, Cicchetti DV. Observation, history, and physical examination in diagnosis of serious illnesses in febrile children = 24 months. J Pediatr 1987; 110 : 26-30. [PUBMED]
18.Baker MD, Avner JR, Bell LM. Failure of infant observation scales in detecting serious illness in febrile, 4-to 8-week- old infants. Pediatrics 1990; 85 : 1040-1043. [PUBMED]
19.Teach SJ, Fleisher GR, Occult Bacteremia Study Group. Efficacy of an observation scale in detecting bacteremia in febrile children three to thirty-six months of age, treated as outpatients. J Pediatr 1995; 126 : 877-881.
20.McCarthy PL, Sznajderman SD, Lustman-Findling K, Baron MA, Fink HD, Czarkowski N et al. Mothers' clinical judgment: a randomized trial of the Acute Illness Observation Scales. J Pediatr 1990; 116 : 200-206.(Bharti Bhavneet, Bharti Sahul, Verma Van)