Evaluation of

pulse oximetry as
a tool to prevent childhood pneumonia
related morbidity and mortality
Study Report

Regional Technical Resource Centre for Health Technology Assessment,

Achutha Menon Centre for Health Science Studies,
Sree Chitra Tirunal Institute for Medical Sciences and Technology,
Trivandrum - 695011

In Collaboration with

Health Technology Assessment in India (HTAIn),

Department of Health Research (DHR), Ministry of Health and Family Welfare

(MoH&FW), Government of India

September 2019
Evaluation of Pulse Oximetry as a
Tool to Prevent Childhood
Pneumonia Related Morbidity and
Mortality

Regional Technical Resource Centre for Health Technology Assessment,

Achutha Menon Centre for Health Science Studies,
Sree Chitra Tirunal Institute for Medical Sciences and Technology,
Trivandrum - 695011

In collaboration with

Department of Health Research (DHR), Ministry of Health and Family Welfare

(MoH&FW), Government of India

September 2019
1
List of Contributors:

Dr. Hisham Moosan, Technical Expert – Epidemiology & Project Lead, Regional
Technical Resource Centre for HTA (RTRC-HTA), Achutha Menon Centre for Health
Science Studies, SCTIMST Trivandrum

Dr. Antony Stanley, Research Associate, RTRC-HTA, Achutha Menon Centre for Health
Science Studies, SCTIMST Trivandrum

Dr. Raman Kutty V., Principal Investigator, RTRC-HTA Chief & Emeritus Professor,
Achutha Menon Centre for Health Science Studies, SCTIMST Trivandrum

Dr. Biju Soman, Co-Principal Investigator, RTRC-HTA & Professor, Achutha Menon
Centre for Health Science Studies, SCTIMST Trivandrum

Ms. Priya Abraham, Data Manager, RTRC-HTA, Achutha Menon Centre for Health
Science Studies, SCTIMST Trivandrum

Dr. Oshima Sachin, Scientist D, HTAIn Secretariat, Department of Health Research,

MoHFW, Govt. of India

Ms. Jyotsna Naik, Scientist C, HTAIn Secretariat, Department of Health Research,

MoHFW, Govt. of India

Dr. Malkeet Singh, Junior Health Economist, HTAIn Secretariat, Department of Health
Research, MoHFW, Govt. of India

Suggested citation:

Moosan H, Stanley A, Raman Kutty V, Soman B, Abraham P. Evaluation of pulse oximetry

as a tool to prevent childhood pneumonia related morbidity and mortality. Regional
Technical Resource Centre for Health Technology Assessment, Achutha Menon Centre for
Health Science Studies, Sree Chitra Tirunal Institute for Medical Sciences and Technology,
Trivandrum and Department of Health Research, Ministry of Health and Family Welfare,
Government of India. 2019

2
Acknowledgements

We would like to extend our sincere gratitude to the Secretaries of DHR, for selecting
SCTIMST, Trivandrum as an HTA Resource Hub, assigning projects and building the
team’s capacity to carry out the work of HTAIn, DHR.

We are also deeply thankful to the Additional Secretary, Joint Secretary and Deputy
Secretaries of the DHR for the continuous administrative support.

We gratefully acknowledge the work of our former Additional Chief Secretary, Sri Rajeev
Sadanandan IAS, for his unwavering commitment to establish a centre for HTA in Kerala.

We would like to thank Dr. Rajan N. Khobragade, Principal Secretary, Health and Family
Welfare, Government of Kerala for his unconditional motivation and support for the
centre and its activities.

We would also like to thank Sri Keshvendra Kumar IAS, State Mission Director, National
Health Mission, Kerala for his continued support and partnership with the Regional
Technical Resource Centre.

We thank the members of the Technical Appraisal Committee, HTA In, for their
constructive suggestions throughout this study.

We sincerely thank Dr. Santhosh Kumar, Professor and Head, Department of Paediatrics
and Dr. Purushothaman K K, Paediatrician, Government Medical College, Thrissur, for
their valuable inputs related to the protocol of diagnosing and management of childhood
pneumonia.

We would also like to thank Dr. Nita Vijayan, State Programme Manager (RCH) for taking
the time to discuss concerns on childhood pneumonia, current standards of prognostic
tools available in India, and the issues around the treatment and length of stay for
children with severe pneumonia.

We express our sincere thanks Dr. Sreehari M. State Nodal Officer, CH & RBSK for his
inputs on the different types of oximetry devices used to treat and monitor different
stages of pneumonia.

3
We would like to thank Dr Malkeet Singh, Junior Health Economist, Department of Health
Research, for his technical support in developing the decision tree model for this study.

We thank Dr Mohammad Ameel, Senior Consultant, Healthcare Technologies (Medical

Devices) at National Health Systems Resource Centre for his inputs on the pulse oximeter
device specification.

We would like to express our heartfelt gratitude to Dr Kavitha Rajsekar, Scientist E, DHR
& HTAIn Co-Ordinator, without whose feedback and constant encouragement this report
could not have been completed. And finally we would like to express our sincere thanks
and acknowledgement to the team at Department of Health Research (DHR) especially
Dr. Oshima Sachin (Scientist D) and Ms. Jyotsna Naik (Scientist C), Government of India,
for their unwavering support and guidance throughout the study.

4
 Table of Contents

Ch.No. Title Page

i Executive Summary 8

1 Introduction 11

2 Methods (Systematic Review) 13

a. Objectives 13

b. The PICO 13

c. Types of studies considered 14

3 Results (Systematic Review) 16

a. Search Results 16

b. Risk of Bias 16

c. Disease Progression 16

d. Care-seeking and health-care parameters 17

e. Prognostic parameters 17
f. Sensitivities and specificities based on thresholds of
18
hypoxemia
g. Cost parameters 20

h. Feasibility, Usability and Acceptability 20

4 Discussion (based on systematic review) 21

5 Supply of oxygen in hospitals 24

Costing of incorporation of Pulse-oximeter to IMCI
6 26
guidelines in Indian primary health care
a. Methodology 26

b. Results 26

7 Budget Impact Analysis 27

a. Methodology 27

5
 b. Results 28

8 Economic Evaluation 29

a. General Model Description 29

b. Sensitivity analysis 31
Discussion (based on decision model and budget impact
9 33
analysis)
10 Conclusions and Recommendations 35

11 Limitations and assumptions 35

12 References 36

13 Addendum 40

6
List of tables

Table 1. Details of PICO used for the study

Table 2. Sensitivity, Specificity, Positive Predictive Value & Negative Predictive
Value of Pulse oximetry to diagnose pneumonia
Table 3. Value of low oxygen saturations in predicting outcome
Table 4. Pulse Oximeter specifications as per the Technical Specifications of
Medical Devices for Neonatal & Paediatric Care ICUs
Table 5. Results of Cost effectiveness of IMCI + PO as compared to IMCI Alone
Table 6: Evidence Summary
Table 7: Summary of studies with supportive evidence for pulse oximetry
Table 8 A: Parameters of Disease Progression
Table 8 B: Care-seeking and health-care parameters
Table 9: Prognostic parameters
Table 10: Cost parameters
Table 11: Annualization formula
Table 12: Cost of treatment
Table 13: Cost of training frontline health-worker
Table 14: Assumptions made for the decision tree in the study

List of figures

Figure 1: PRISMA Chart

Figure 2: Decision Model – IMCI
Figure 3: Decision Model – IMCI + PO
Figure 4. Tornado diagram for one-way sensitivity analysis
Figure 5. Results of PSA plotted against Base-Case ICER
Figure 6: Cost Effectiveness Acceptability Curve

7
Executive Summary

Introduction:

Globally, pneumonia is the leading cause of death in children <5 years of age. Despite
interventions being available, it is estimated that pneumonia is responsible for 15% of
childhood deaths worldwide. In the absence of appropriate prognostic tools at the
frontline, currently recommended World Health Organization (WHO) guidelines for
integrated management of childhood illness (IMCI) often lead to an overuse of antibiotics
and the under-referral of patients with severe pneumonia who require hospital care.

The present recommended strategy for diagnosis and prognosis of pneumonia is IMCI
tool for professional health workers at health facilities and iCCM tool for community
health workers. Currently, identification of these IMCI symptoms remains inconsistent
and unreliable among health-care personnel. The use of pulse-oximetry devices (used to
measure the oxygen level in the blood) in community health-care settings has been
proposed as a method to identify hypoxic children at risk of treatment failure.

Objectives:

 To determine the effectiveness (i.e. sensitivity, specificity, positive and negative

predictive values) and cost effectiveness of pulse oximetry devices in screening of
childhood pneumonia by health workers in resource poor settings (LMICs).
 To identify whether children have lower mortality rates, lower morbidity, and
shorter length of stay where pulse oximeters are used to inform diagnosis and
treatment compared with where pulse oximeters are not used.

Methods:

A systematic literature review (SLR) was conducted with PICO as population consisting
of patients aged 0-5 years with pneumonia in LMIC’s, Intervention was IMCI +Pulse
Oximetry and the comparator was IMCI alone. The outcomes was diagnostic accuracy of

8
pulse oximetry, cost per QALY gained, and incremental deaths averted with the
introduction of pulse oximetry. The SLR was conducted using electronic databases such
as Database of Cochrane, PubMed, Web of Science, ProQuest, Google Scholar and WHO
Global Health Library, with no language restrictions. A decision tree model was
developed for estimating the Incremental Cost Effectiveness Ratio (ICER). For populating
the model we required data pertaining to disease progression, care-seeking, health-care,
prognostic and cost parameters. In addition to data gathered from SLR, we did a scoping
review wherever necessary to extract data for the decision model.

We extracted the costs of the IMCI implementation in Indian primary healthcare settings
from previous studies. The average number of hospital visits (outpatient and inpatient)
were calculated including the associated out of pocket health expenditure for each
category. We indexed the amount to the current INR rate using inflation tools. Pulse
oximeters which fit the predefined specifications were shortlisted and their net average
cost was taken for costing purposes. The training cost is derived from consultation with
the program implementers.

Results:

The evidence from the systematic review was overwhelmingly in favour of the use of
pulse oximetry along with the existing guidelines. Out of the seven studies which were
eventually shortlisted, 6 of them favoured the use of PO. When it came to costing, the cost
of Pulse oximetry per patient per year is INR 0.36. We did a budget impact analysis, which
showed that if we were to provide a pulse oximeter to all the PHC’s in India, the cost of
the roll-out of pulse oximeters would be INR 64,125,000. The cost of training frontline
health workers to use PO is INR 26,334,000. The overall cost of roll-out of pulse oximeters
in PHC’s would amount to INR 90,459,000. In absolute terms, the introduction of pulse-
oximetry devices to IMCI is estimated to result in annual reductions in pneumonia deaths
in India. The deaths averted per year for PO2 within a cohort of 10000 pneumonia cases
would be 367 if the sensitivity is 85% and 213 if the sensitivity is 70%. Owing to the large
under-five populations (128 million) India could significantly reduce the mortality due to
childhood pneumonia by the introduction of PO into the existing IMCI.

9
Recommendations:

 IMCI+PO is a cost saving prognostic tool as compared to IMCI alone provided there
is supplementary oxygen availability.
 IMCI should be the basic prognostic tool for childhood pneumonia but PO is
beneficial in the referral of cases.
 Pulse oximetry in general may be used to measure oxygen saturation in cases
wherever required. Among outpatients with pneumonia, oxygen saturations
<90% were associated with increased morbidity and mortality.
 A hospital admission threshold of <92% would be safer and clinically better
justified.
 All severe cases, irrespective of availability of Pulse oximeter, should be referred
to a tertiary care facility for expert management.

10
 Evaluation of pulse oximetry as a tool to prevent
childhood pneumonia related morbidity and mortality

1. Introduction

Pneumonia is an infection of the lower respiratory tract that involves the airways and
parenchyma with consolidation of the alveolar spaces. Lobar pneumonia describes
pneumonia localized to one or more lobes of the lung. Atypical pneumonia describes
patterns typically more diffuse or interstitial than lobar pneumonia. Infectious agents
that commonly cause community-acquired pneumonia vary by age1. Most common
causes are respiratory syncytial virus (RSV) in infants, other respiratory viruses
(parainfluenza viruses, influenza viruses, human meta-pneumovirus, adenoviruses) in
children younger than 5 years old, and Mycoplasma pneumonia in children older than age
5 years. Streptococcus pneumonia is the most common bacterial cause of lobar
pneumonia, and occurs in children of any age outside the neonatal period1. M.
pneumoniae and Chlamydophila pneumoniae are principal causes of atypical pneumonia.

Globally, pneumonia is the leading cause of death in children <5 years of age 2. Despite
interventions being available, it is estimated that pneumonia is responsible for 15% of
childhood deaths worldwide. Reductions in annual mortality remain modest, with nearly
950,000 under-5 year olds dying of pneumonia in 2013. Immunizations have brought
down the incidence of pneumonia caused by Haemophilus influenza type b, and S.
pneumoniae. But despite the unprecedented rate of Haemophilus influenzae type B (Hib)
and pneumococcal vaccine (PCV) introduction, achieving high levels of coverage in
developing countries is still challenging. Therefore, in regions where vaccine introduction
and scale-up lags behind other countries, improved access to diagnosis and treatment is
crucial.

A crucial component of improving pneumonia outcomes is the early identification of

patients at risk of treatment failure and the timely provision of supportive care 3.
However, in the absence of appropriate prognostic tools at the frontline, currently
recommended World Health Organization (WHO) guidelines for integrated management

11
of childhood illness (IMCI) often lead to an overuse of antibiotics and the under-referral
of patients with severe pneumonia who require hospital care4,5.

IMCI is a global strategy designed to strengthen health systems, implement clinical

guidelines in health facilities, and implement community based interventions that might
include care by community health workers (CHW). By using simple diagnostic algorithms
based on a small number of clinical signs CHWs can detect and treat diseases like
pneumonia to some extent. The most recent 2015 technical update of IMCI guidelines
defines non-severe pneumonia as the presence of fast breathing or chest in-drawing or
both, which is treatable with oral antibiotics. Severe pneumonia is defined as cough or
difficulty breathing in the presence of danger signs, and requires referral to a hospital or
health facility for injectable antibiotics or other supportive care such as oxygen therapy.
Currently, identification of these IMCI symptoms remains inconsistent and unreliable
among community health-care workers or carers without clinical training. Therefore,
improved prognostic and diagnostic tools for case-management are necessary to
substantially reduce pneumonia-associated morbidity and mortality6. Hypoxaemia and
malnutrition are strong predictors of mortality in children who are hospitalized for
pneumonia. This has led to increasing support for the use of oxygen therapy and
monitoring oxygen saturation in the management of severe cases. It is estimated that
15% of children who are hospitalized for pneumonia have hypoxaemia7 (oxygen
saturation, or SpO2, of <90%) and that around 1.5 million children with severe
pneumonia require oxygen treatment each year. Misclassification of a disease as
pneumonia can also lead to the potential overuse of Amoxicillin thereby leading to
antibiotic resistance.

The use of pulse-oximetry devices (used to measure the oxygen level in the blood) in
community health-care settings has been proposed as a method to identify hypoxic
children at risk of treatment failure. Unfortunately, current treatment coverage remains
low, and, more importantly, most childhood pneumonia deaths result from a lack of, or
delay in, accurate diagnosis. These devices may be particularly beneficial at the frontline
given that they require little training and reduce the reliance on clinical symptoms. The
current pulse-oximetry systems are also quick, non-invasive and require minimal
infrastructure. As international organisations are investing in programmes to increase
pulse oximeter use in low-income settings, more research is needed on the optimal use

12
of pulse oximeters (eg, appropriate oxygen saturation thresholds), and how pulse
oximeter use affects referral and admission rates, length of stay, resource utilisation and
health outcomes. We therefore reviewed the evidence on the effectiveness, and cost-
effectiveness of pulse oximetry devices introduction alone and when combined with the
existing IMNCI guidelines.

2. Methods:

a. Objectives

This study attempted to address the following questions:

c. “What is the effectiveness (i.e. sensitivity, specificity, positive and negative

predictive values) and cost effectiveness of pulse oximetry devices in screening of
childhood pneumonia by health workers in resource poor settings (LMICs)?”
d. “Do children have lower mortality rates, lower morbidity, and shorter length of
stay where pulse oximeters are used to inform diagnosis and treatment compared
with where pulse oximeters are not used?”

b. The PICO

Table 1. Details of PICO used for the study

Population
 Patients aged 0-5 years with pneumonia
Intervention
 IMCI + Pulse Oximetry
Comparator
 IMCI
Outcome
 Diagnostic Accuracy of Pulse Oximetry
o (Sensitivity, Specificity, Positive and Negative Predictive Values)
 Referral and admission rates, length of stay, resource utilisation and health
outcomes
 Cost per QALY gained/ DALY averted
 Incremental deaths averted with the introduction of pulse oximetry by
community health workers in community settings

13
c. Types of studies considered

The types of studies considered for the review included the following:

For the effectiveness we considered randomized controlled trials, field/ community trials
and quasi-randomized trials. The cost effectiveness data was extracted from full
economic evaluations (studies in which both the costs and outcomes of the alternatives
were examined and in which a comparison of two or more interventions or case
alternative was undertaken) including trial-based, non-trial based (i.e. observational
studies), simulation-based, decision model, trial-based model and partial economic
evaluations such as cost analysis, costing of interventions with or without comparator.

The population selected were patients aged 0-5 years with pneumonia in LMIC’s.
Intervention was the use of Pulse oximetry alone or in combination with the existing
IMNCI guidelines. The comparators were IMCI alone or the existing local pneumonia
management guideline. In other words, we included studies with at least one intervention
group in which a pulse oximeter reading was taken and at least one control group in
which pulse oximetry was not used.

The outcomes of interest included diagnostic accuracy of pulse oximetry, referral and
admission rates, and length of stay, resource utilisation and health outcomes, cost per
QALY gained/ DALY averted, incremental deaths averted, difference in mortality and
morbidity, referral rates and length of stay in critical care where pulse oximeters are used
to inform diagnosis and treatment compared with where pulse oximeters are not used.

We systematically searched the Database of Cochrane, PubMed, Web of Science,

ProQuest, Google Scholar and WHO Global Health Library, with no language restrictions.
Study references were checked. Websites of non-governmental organisations, health
organisations and development organisations were searched for unpublished reports
using the terms ‘pulse oximeter’ and ‘pulse oximetry’. Topic experts were contacted for
additional materials. Studies that were not relevant based on title/abstract were
excluded. We read the remaining studies’ full texts and excluded those not fulfilling the
inclusion criteria. All full texts were read by a second person, and if inclusion was
uncertain, a third. We extracted data using a tailored Cochrane data collection form and
assessed risk of bias using the Cochrane ACROBAT tool8. We intended to calculate risk
ratios, mean differences and CI’s, and if possible pool data within subgroups and conduct

14
a meta-analysis. However, due to the small number of studies and the study
design/outcome variability this was not possible. Instead we narratively describe the
evidence using a structured approach while drawing insights where possible, a standard
strategy in such situations.

Figure 1: PRISMA Chart

Studies obtained from

database searches (PubMed,
Web of Science, Cochrane
Library and Google Scholar -
361

254 studies after duplicates 107

were removed duplicates
removed

Inclusion/Exclusion
criteria applied

204 studies
removed
50 studies after screening of
based on
Title & Abstract
incl/excl
criteria

Full Paper Review

13 studies having information 30 studies summarily rejected

related to due to non-relevance
feasibility/acceptability/usabi
lity/impact

7 studies having information

relevant to effectiveness
and/or cost effectiveness
included in the main review

15
3. Results of the Systematic Review

a. Search Results

We found 361 reports initially, which underwent removal of duplicates and application
of the inclusion and criteria, followed by screening of all titles and abstracts. The
shortlisted potentially relevant studies amounted to 50. They were divided into three
categories based on their relevance to the PICO. Studies with serious and critical level of
risk of bias were not included in the final summary. Most of the studies fell into the
moderate risk category. Seven studies, were closely aligned (see PRISMA flow diagram)
to the PICO, and thirteen studies were relatively associated with the PICO. Finally 20
studies were considered for inclusion. We’ve added them to the analysis as supportive
evidence (see Table 2).

b. Risk of Bias

Table 6 and 7 in the addendum of this document demonstrates the risk of bias for each
study included.

c. Disease Progression

In 2013, Rudan et al.9 estimated that in India alone, the number of new episodes
(incidence) of community–acquired pneumonia in children 0–4 years of age is
127,960,004. The number of new severe episodes (according to WHO's definition) that
required hospitalizations was 4,066,541. The estimates of the number of child deaths
attributable to pneumonia was 388,144.

By 2015, Farooqui et al.10 projected that there were 3.6 million (3.3–3.9 million) episodes
of severe pneumonia and 0.35 million (0.31–0.40 million) all cause pneumonia deaths in
children younger than 5 years in India. In addition, to the overall figure, state-wise
breakup of data suggested that the states that merit special mention include Uttar
Pradesh where 18.1% children reside but contribute 24% of pneumonia cases and 26%
pneumonia deaths, Bihar (11.3% children, 16% cases, 22% deaths) Madhya Pradesh
(6.6% children, 9% cases, 12% deaths), and Rajasthan (6.6% children, 8% cases, 11%
deaths). In the latest Pneumonia and Diarrhoea Progress Report 2018, the number of
Pneumonia deaths in children under 5 years has come down to 158,176.

16
The proportion of pneumonia severe on day 1 was 5% (2–10%) according to Pitt et al.11
The mean duration of non-severe illness before recovery was 3 days (2–4 days)12. The
mean duration of non-severe illness before progression to severe illness was 10 days (9–
11 days)12. The mean duration of severe illness before recovery was 4 days (3–5 days)13.
Mean duration of severe illness before death was 7 days (6–8 days)14. The proportion of
bacterial versus viral infection was 15% bacterial (10-25%) and 85% viral (75–90%)15.

d. Care-seeking and health-care parameters

The mean duration of illness before care seeking in the case of non-severe illness was 3
(2–4) days and 0.75 (0.5–1) days in the case of severe illness14. Foran M et al.16 found
that oximetry data changed clinical management in all observed cases of hypoxemia and
several cases of normoxemia, leading to application of supplemental oxygen, initiation of
further diagnostic testing, prolongation of inpatient stay, or expedited discharge home.
According to McCollum et al., the availability of oximetry appeared to have increased the
referral rate for severely hypoxaemic children without chest indrawing or danger signs
from 0% to 27.2%. In the absence of oximetry, if the relevant World Health Organization
(WHO) guidelines published in 2014 had been applied, 390/568 (68.7%) severely
hypoxaemic children at study health centres and 52/84 (61.9%) severely hypoxaemic
children seen by community health workers would have been considered ineligible for
referral17,18.

The Pneumonia & Diarrhea Progress Report 2018 states that in India, 73% of patients
with pneumonia were taken to an appropriate healthcare provider. The probability of
community-based treatment curing non-severe bacterial case was 0.925 (0.90–0.95)19
and the probability of treatment with amoxicillin curing severe case if it adhered to the
prescription (based on treatment failure rates of patients with hypoxia at baseline) was
0.65 (0.6–0.7). However the optimal dosing recommendation for amoxicillin remains
unclear with limited pharmacological and clinical evidence20.

e. Prognostic parameters

Clinical signs alone or in combination are not suitable to diagnose hypoxaemia 21.
According to Fu LY et al.22 the inclusion of oximetry data improved the predictive ability
at baseline, 12 hours, and 24 hours. The ability to predict failure after 12 hours of

17
observation with oximetry data was similar to the predictive ability after 24 hours
without pulse oximetry data.

The sensitivity of IMCI was 0.55 (0.5–0.6)23 whereas the sensitivity of pulse oximetry
with IMCI was 0.85 (0.8–0.9)24. Pulse oximetry misclassified notably fewer well children
than did the WHO algorithm (4% vs 35%)24. Pulse oximetry and the WHO algorithm
together (SATWHO) detected 99% and 87% of pneumonic ALRI and radiologic
pneumonias, respectively, and both methods detected 94% of all cases of pneumonic and
non-pneumonic ALRI diagnosed clinically. No single clinical sign can perform as well as
pulse oximetry for predicting hypoxia in children with severe pneumonia25.

The adherence of community health workers to management guidelines especially the

dosing and referral guidelines in non-severe prognosis (IMCI) is 0.55 (0.5–0.6)26 and 0.65
(0.6–0.7) in case of severe prognosis (IMCI). The general perception is that a physical tool
would increase the likelihood of adherence, hence for pulse oximetry use it is set at a
range of 65% to 85%.

f. Sensitivities and specificities based on thresholds of hypoxaemia27

Hypoxaemia definitions varied from <92% to 95%. According to Majumdar et al., raising
the admission threshold to 92% entails 1 additional hospitalization for every 14 patients
discharged. Among outpatients with pneumonia, oxygen saturations<90% were
associated with increased morbidity and mortality28. Their results indicated a hospital
admission threshold of <92% to be safer and clinically better justified.

Table 2. Sensitivity, Specificity, Positive Predictive Value & Negative Predictive

Value of Pulse oximetry to diagnose pneumonia

Saturation N Died or OR Sensitivity Specificity PPV NPV

level Critical (95% (%) (%) (%) (%)
Care CI)
(%)
SpO2 (n=69) 36.2 3.3 29.8 88.5 36.2 85.2
<88% (1.9-
5.7)

18
 SpO2 (n=131) 29.8 2.7 46.4 76.0 29.8 86.6
≤90% (1.7-
4.5)
SpO2 (n=187) 26.7 2.6 59.5 64.2 26.7 87.9
≤92% (1.6-
4.3)
SpO2 (n=271) 22.5 2.2 72.6 45.2 22.5 88.3
<95% (1.3-
3.7)
** Adapted from Thomas Bewick et al. 2010

Table 3. Value of low oxygen saturations in predicting outcome27

SpO2 >90% (n=336) SpO2 ≤90% (n=131)

Inpatient death by 30 6.8 19.1
days (%)
Critical care admission 7.4 15.3
(%)
Mechanical ventilation 1.8 5.3
Median LOS in days (IQR) 6.56 (8.54) 9.75 (10.33)

Among outpatients with pneumonia, oxygen saturations <90% were associated with
increased morbidity and mortality. A hospital admission threshold of <92% would be
safer and clinically better justified7,28.

Comparison of Arterial Oxygen Saturation Measured Both by Pulse Oximeter

and Arterial Blood Gas Analyzer in Hypoxemic and Non-hypoxemic Pulmonary
Diseases29
In pulmonary diseases with SpO2 ≥ 80%, pulse oximetry has high accuracy in
estimating SaO2 and may be used instead of arterial blood gases (ABG). In
patients with SpO2 < 80%, however, the exact estimation of SaO2 and the
evaluation of oxygenation by pulse oximeter is not a good substitution for ABG
analyzer.

19
g. Cost parameters

The cost of Amoxicillin treatment per child is INR 47.77 as per Jan Aushadi rates
(http://janaushadhi.gov.in/old-data/New_MRP_Oct15.pdf). The average hospital cost
per episode is INR 207 if it’s a primary healthcare facility, INR 293 when it is a secondary
healthcare facility and INR 437 when the referral is to a tertiary healthcare facility (WHO
CHOICE database). The cost of a USFDA approved Life Box Pulse oximeter is $250. 30 Each
set of batteries can be used for approximately 840 readings. The lifetime of device is 2
years.31 There are ISO certified finger pulse oximeters available in the GeM which cost
anywhere between INR 1800 – 4000. The number of devices needed is 1 per 1,000
children under 5 (0.8–1.2). The DALY’s lost due to pneumonia32 in young children amount
to 121·15 (105·92 to 138·3) * 10^5.

Floyd et al.,33 evaluated the incremental cost-effectiveness of pulse oximetry in

comparison to a baseline of using IMCI alone. The costing was undertaken from a public
health provider perspective and hence no societal, economic or private sector costs were
included. Since sufficient data were not available to inform the sensitivity of IMCI when
combined with pulse oximetry, two scenarios were proposed: one in which the addition
of pulse oximetry increases the sensitivity of IMCI to 70%, and one in which the
sensitivity of the combination is increased to 85%, reflecting the potential of pulse
oximetry to identify both people with hypoxic cases and cases with abnormal oxygen
saturation (90–95%) who would benefit from referral. Estimated cost-effectiveness (US$
per disability adjusted life year (DALY)) of combined integrated management of
childhood illness (IMCI) and pulse oximetry with 70% sensitivity and combined IMCI and
pulse oximetry with 85% sensitivity compared with IMCI were US$14 and US$12 per
disability-adjusted life year averted respectively.

h. Feasibility, Usability and Acceptability

Ginsburg AS et al., reported that healthcare professionals and caregivers viewed the pulse
oximeter and breath counter favourably. Challenges included electricity requirements for
charging and the time needed to complete the application. Some caregivers saw it as a
sign of modernity, increasing their trust in the care received. Other caregivers were
hesitant or confused about the new technology. Overall, this technology was valued by
users and is a promising innovation for improving quality of care in frontline health

20
facilities34. In Spence et al., community health workers (CHW) and national stakeholders
across the four countries perceived the acute respiratory infection (ARI) timer and
fingertip pulse oximeter as highly scalable and easy for CHWs to use. CHWs placed greater
priority on device acceptability to caregivers and children. Both groups felt that heavy
reliance on electricity reduced potential scalability and usability in rural areas. Device
simplicity, affordability and sustainability were universally valued35.

4. Discussion (based on the systematic review)

In general, the states having a high prevalence of pneumonia risk factors and poor access
to health services had a higher burden of pneumonia cases and deaths10. Although
hypoxemia is common, the absence of routine pulse oximetry results in most hospitalized
hypoxemic children not receiving available oxygen treatment. It is important to recognize
that referral/ admission rates is dependent on the thresholds for oxygen therapy. Setting
a uniform threshold for a country as geographically varied as India is complicated
because pulse oximeter results may not be considered in isolation from clinical findings,
SaO2 can naturally fluctuate over a day, and studies show that ‘healthy’ SaO2 differs by
age and altitude.

Clinical signs alone are poor predictors of hypoxemia, and using pulse oximetry in
resource-poor health facilities to target oxygen therapy is likely to save costs. The relative
ease of implementation of a pulse oximetry-based intervention (even with the
assumption of perfect availability) compared with the elimination of low birth weight or
malnutrition makes it an important candidate for an intervention against pneumonia in
resource-poor settings. Many of these studies considered have not taken into account the
positive impact pneumococcal vaccine would have for infants. In Sinha et al. 13, it is seen
that the vaccine has the potential to directly avert around 262,000 deaths in under-5s
across 72 countries. Taking that vaccine into consideration might decrease the cost-
effectiveness of pulse oximetry. Still it should compare favourably with the pneumococcal
vaccine.

The breadth of different views on pneumonia diagnostic aids emerging from the research
activities can be seen amongst the different participant groups. The fingertip pulse
oximeter was deemed the most usable and scalable in most of the studies. There is
evidence to indicate that pulse oximetry may lead to improved health outcomes, with

21
lower mortality rates (when combined with improved/adequate oxygen administration);
pulse oximetry may change physicians’ decisions regarding illness severity, and increase
hospital admissions related to previously unrecognised hypoxaemia. Routine pulse
oximetry may also influence diagnostic tests and treatments used. Pulse oximetry could
facilitate swifter diagnosis, so effective treatment starts earlier and recovery likelihood
increases, reducing future resource use. Oximetry can reduce resource waste by
indicating when to end treatment, and by decreasing false positives. The emphasis should
be to ensure all aids are accurate, affordable and acceptable to the communities where
they will be used.

Table 4. Pulse Oximeter specifications as per the Technical Specifications of

Medical Devices for Neonatal & Pediatric Care ICUs36

Weight (gram) Less than 300 g

Noise Level (dB) <40 dB

GMDN name and Code Pulse oximeter CT 1446

Definition A photoelectric device intended for the
continuous transcutaneous measurement and
display of haemoglobin oxygen saturation
(SpO2). The signals, typically produced by light-
emitting diodes (LEDs) and a receiving detector
in a probe, or directly built-in, are used to make
the measurements using the principle of
spectrophotometry The oximeter displays the
SpO2 values and may calculate / display other
parameters, e.g., pulse rate, respiratory rate.
Technical characteristics a. SpO2 measurement range at least 40-70 and 70
(specific to this type of device) to 99 %, minimum gradation 1%
b. Accuracy of SpO2 better than +-1% for range
40-70 and better than +-3% for range 70-99
c. Accuracy of pulse rate better than ± 5 bpm

22
 d. Audio-visual alarms required: high and low
SpO2 and pulse rate; operator variable settings;
sensor disconnected, sensor failure, low battery
e. Pulse rate range at least 30 to 240 bpm,
minimum gradation 1 bpm
Power consumption 1.5 Watt
Power input and frequency 220 to 240V, 50 Hz
Display type Color OLED

Heat Dissipation Should maintain nominal temperature and

prevent overheating of the probe
Mobility, portability Protective splash-proof case for clean storage
and safe transport
Operating condition Capable of operating continuously in ambient
temperature of 0 to 50 degree Celsius and
relative humidity of 15 to 90% in ideal
circumstances
Standards & Safety Should be FDA / CE approved product / ISO
80601-2-61-2011: Medical Electrical equipment-
part 2-61: Particular requirements for the basic
safety and essential performance of pulse
oximeter.

Manufacturer/ supplier should have ISO 13485

certificate for quality standard.

Electrical safety conforms to standards for

electrical safety IEC-60601-1, shall meet IEC-
60601-1-2 (General requirements for safety -
electromagnetic compatibility)
Signal strength or quality to Yes
be visually displayed

23
 Memory Minimum 24 hour trend memory for SpO2 & PR

User's interface Easily accessible touch button to operate the

machine
Display must allow easy Yes
viewing in all ambient light
levels
Warranty 1 year warranty with minimum 3 free servicing.

Accessories Sell in standard:

-A lithium battery
-A user manual
-A USB line
-A New born oximeter probe
-A power adapter

Sell in addition:
-Adult prone
-Pediatric probe

5. Supply of oxygen in hospitals

The Public Health Standards of India which was published in 2012 detailed the minimum
expected infrastructure and human resources needed to run the country’s healthcare
services, from sub-centres to district hospitals. These standards apply to government-run
healthcare facilities. As per the document, oxygen is an essential commodity for any
maternal birthing unit, including sub-centres, the most ‘peripheral’ of India’s rural health
facilities, and that all district hospitals should have piped oxygen to their sickest patients
(intensive care units and casualty). It is important to make sure that the procurement is
done by appropriately by fairly awarding contracts, payments without delay, and
prevention of theft of procured stocks. The challenge is to create a system which makes
these outcomes more likely and limits harm to patients from almost inevitable human
and other error.

Oxygen therapy must be more widely available; in many remote settings, this can be
achieved by use of oxygen concentrators, which can run on regular or alternative sources
of power. Several conditions must be met for hypoxemic children to receive appropriate,

24
uninterrupted oxygen therapy for as long as is necessary to save their lives. A 2016 gap
analysis of gas use in the Armed Forces Medical College Hospital in Pune found limited
documentation from the oxygen supplier and major gaps in pipeline maintenance,
particularly in the lack of alarm mechanisms to alert staff to shortages. Medical oxygen
comes under the National List of Essential Medicines (NLEM) of India, one of the key
instruments in a balanced healthcare delivery system. Medical oxygen is also on the
World Health Organization’s (WHO) list of essential medicines. No ICU or hospital for that
matter can run without a smooth supply of oxygen as medical oxygen comes under
essential medical list.

25
6. Costing of incorporation of Pulse-oximeter to IMCI guidelines in
Indian primary healthcare

a. Methodology

We extracted the costs of the IMCI implementation in Indian Primary healthcare settings
from previous studies37. The average number of hospital visits (outpatient and inpatient)
were calculated including the associated out of pocket health expenditure for each
category. We indexed the amount to the 2019 INR rate using inflation tools. Pulse
oximeters which fit the predefined specifications were shortlisted and their net average
cost was taken for costing purposes. The cost for the equipment (pulse oximeters) were
derived from the Government e-Marketplace. The power consumption for the finger
pulse oximeter was negligible. The training cost is derived from consultation with the
program implementers. We assumed the life of a training to be two year (similar to the
lifespan of the device) and therefore dealt with the training cost as capital in nature. The
life of training was considered to be 2 years, based on a need for retraining health
workers after an interval of 2 years. Since the PO’s will be handled by the frontline health
workers, no additional human resources will be required to implement this project.

b. Results

Per patient cost of treatment of non-severe pneumonia (home based) before recovery
came to INR 1117 (Table 12). This figure was arrived at by assuming that the patient had
to make at least 1 OPD Visit and the associated OOP. Per patient cost of treatment of non-
severe pneumonia (home based) before progression to severe illness came to INR 2234
taking into account 2 OPD visits and OOP. Per patient cost of treatment for severe
pneumonia (hospital based) is INR 10592. Per infant cost of general health system
administration and program cost (IMCI) came to INR 16137. The average cost of a pulse
oximeter in GeM portal with the required specification is INR 2500. The cost of training
one frontline health worker is INR 513.33 (Table 13). This takes into account the cost of
trainer honorarium, stationery, and reimbursements for trainee, infrastructure and
logistics, and travelling costs. We want to train two frontline health worker per facility.
Hence the cost per facility will be INR 1026.66. Summing up the common expenses for
training two health workers per facility and the instrument cost, the total cost to install
pulse oximeter in a PHC will be INR 3526.67. The expected life of the equipment is 2 years.

26
The annualization factor was taken 0.5226 and the discount rate was 3% (Table 11).
Therefore, per year cost of PO will come to INR 1833.52 (3526*0.522). The number of
patients in the 0-5 age group with acute respiratory infection visiting a PHC OPD (other
than immunisation) in a day is around 10-15, which may increase up to 25 depending on
the availability of the paediatrician. This indicator when extrapolated to a year comes to
around 5040. The cost of Pulse oximetry per patient per year is INR 0.36

7. Budget Impact Analysis

Since the pulse oximeter devices are being planned to be rolled out at a national level in
a phased manner, it is all the more prudent to look at the budget impact analysis along
with the cost effectiveness analysis. In contrast to cost-effectiveness analysis, which
measures both cost and clinical outcomes without regards to underlying disease
prevalence, budget impact models focus exclusively on cost and adjust for the
underlying prevalence of disease. Adjusting for disease prevalence is critical, because a
medical technology that is ‘cost-effective’ might apply to a large subset of patients, and
therefore have a prohibitive budget impact. It is important for the health ministry to be
made aware of the overall budget needed for rolling out the pulse oximeter to all the 1.5
lakh Health and Wellness Centres in India. Depending on the overall budget, structured
plans can be made as to whether the roll-out is made in a single phase or in multiple
phases.

a. Methodology

The Ayushman Bharat targets operationalizing 1.5 lakh Health and Wellness Centres in
India. Each centre is expected to have at least one functioning pulse oximeter and two
trained staff. The cost of PO was gathered from the finger pulse oximeters available in the
GeM portal. The net average cost of all the oximeters which had the recommended
specifications were taken. This was multiplied by the number of proposed Health and
Wellness Centres. One year warranty and three free servicing was included in the
recommended terms and conditions. Training of frontline health workers formed
another major part of the budget. The training was planned in a phased manner. Costs for
central level, state level and district level training for trainers was budgeted. This was
followed by training of frontline health workers at the CHC/ PHC levels. The common

27
expenses for the training included trainer honorarium, stationary items, trainee
reimbursement (TA+DA), infrastructure and logistics.

The health system already has a functioning IMCI unit. The pulse oximetry is an add-on
to the existing IMCI. Hence, no additional human resources will be required to implement
it.

b. Results

The cost of roll-out of pulse oximeters for 1.5 lakh health and wellness centres
 Number of health and wellness centres: 150000
 Cost of PO: 2500
 Overall cost of PO: 150000*2500 = INR 375,000,000
But, as on 31st March 2017, there were only 25650 Primary Health Centres (PHCs)
functioning in India38. If we were to provide a pulse oximeter to all the PHC’s in India, the
cost of the roll-out of pulse oximeters would be INR 64,125,000 (25650*2500). The cost
of training frontline health workers to use PO is INR 26,334,000. The overall cost of roll-
out of pulse oximeters in PHC’s would amount to INR 90,459,000.

The number of functioning Community Health Centres (CHCs) in India was 5510 as on
31st March, 201639. Each community health centre would require at least 4 finger-tip
pulse oximeters which should ideally be placed in the casualty, OP and inpatient ward. If
there are dedicated structured pulmonary rehabilitation programme like SWAAS40 an
additional pulse oximeter might be needed in the specific clinic as well. The cost of
equipping all the CHC’s with specified number of pulse oximeters amount to INR
55100000 (5510*2500*4).

28
8. Economic Evaluation

a. General Model Description

A decision tree was parameterized on MS-Excel spreadsheet to estimate the incremental

cost effectiveness of implementing pulse oximetry + IMCI program over IMCI alone.
(Supplementary Material).

A lifetime study horizon starting from base year of 2019 was considered appropriate to
cover all costs and effects comprehensively. Pneumonia have a predominant risk within
the first 5 years of life, with gradually declining risk till about 15 years.

We evaluated the costs and effects from both health system and societal perspective.
Effect was measured in terms of illness episodes averted, child deaths prevented, life
years gained and quality-adjusted life years (QALY) gained. Costs were discounted at 3%
and an annualization factor of 0.52 was added for time preference of cost. We estimated
the standardized unit cost from health system and societal perspective. We report our
findings as incremental cost of implementing IMCI + PO for infants per QALY gained and
per infant death averted as compared to IMCI alone.

The cost data for analysis was extracted from previous costing studies done with IMNCI
and pulse oximetry33,37. The parameters used for the economic model are shown in the
input parameter table. The sensitivity of ‘IMCI’ and the sensitivity of ‘IMCI + PO’ were
taken from previous studies23,24. We started with an initial base cohort of pneumonia
patients of same size for both interventions. Using the abovementioned sensitivity for
IMCI and IMCI+PO respectively, we screened both groups of severe and non-severe
pneumonia. Per patient cost of treatment of non-severe pneumonia (home based) before
recovery, per patient cost of treatment of non-severe pneumonia (home based) before
progression to severe illness, and per patient cost of treatment for severe pneumonia
(hospital based) were calculated from previous studies conducted in India 37. The mean
duration of non-severe pneumonia before recovery, the mean duration of non-severe
illness before progression to severe illness, the mean duration of severe illness before
recovery and the mean duration of severe illness before death were computed from
secondary data12,14.

29
 Table 5. Results of Cost effectiveness of IMCI + PO as compared to IMCI Alone
Intervention Cost LY QALY ICER
IMCI (0.55) 32247526 601736.1 601736.1 -117.32
IMCI+PO (0.85) 29503112 625127.9 625127.9
Intervention Cost LY QALY ICER
IMCI (0.55) 32247526 601736.1 601736.1 -18.7521
IMCI+PO (0.7) 31993096.31 615304.17 615304.17
*Number given in brackets in the first column is the sensitivity of each intervention

In absolute terms, the introduction of pulse-oximetry devices to IMCI is estimated to

result in annual reductions in pneumonia deaths in India. The deaths averted per year for
PO2 within a cohort of 10000 pneumonia cases would be 367 if the sensitivity is 85% and
213 if the sensitivity is 70%. Owing to the large under-five populations (128 million)
India could significantly reduce the mortality due to childhood pneumonia by the
introduction of PO into the existing IMCI. The detailed decision tree model with
probabilities and cost for IMCI and IMCI+PO is given as supplementary material. We
found that the ICER is -117.32 and -18.75 when the sensitivity of the IMCI+PO is 85% and
70% respectively. This suggests that it is a cost-saving intervention from a societal
perspective. The sizeable number of people with non-severe pneumonia who are
undiagnosed develops severe pneumonia and require inpatient treatment in hospitals
increasing the out of pocket health expenditure and the overall cost of treatment.

Figure 2: Decision Model – IMCI

30
 Figure 3: Decision Model – IMCI + PO

b. Sensitivity Analysis
We did one-way sensitivity analysis by varying costs of all input parameters from their
lower limit to upper limit and ascertained their effect on ICER of intervention. We finally
selected input parameters which had maximum impact on cost-effectiveness of
introduction of Pulse-oximeter in IMCI guidelines. We plotted these parameters and
variation in ICER attributed to them in tornado chart as given in Figure 4.

Figure 4. Tornado diagram for one-way sensitivity analysis

Sensitivity of IMCI -157

Sensitivity of IMCI + PO -158 -81

Cure rate of non-severe pneumonia (community-
-148 -86
based)
Probability of non-specific TT with amoxicillin -141 -99

Cost of IMCI+PO -135 -100

Cost of IMCI Only -134 -100

Cure rate of severe pneumonia (hospital based) -122 -95

-160 -140 -120 -100 -80 -60 -40 -20

ICER with Lower Limit ICER with Upper Limit

31
We also conducted a probabilistic sensitivity analysis to ascertain the variation in ICER
which may arise from uncertainty in input parameters and all assumptions we took
during the process of this evaluation. In Microsoft excel, we used Visual Basics to run PSA
where ICER was calculated for 1000 times by randomly varying the values of input
parameters, taking any random value from their lower to upper bound.

All these 1000 iterations were plotted against base case to present the variation and
proportions of ICER falling in different quadrants. We can see, that all values fall in right
upper and right lower quadrant; out of which majority (97.4) of values are in right lower
quadrant, signifying ICER to be negative with gain in QALYs and less cost incurred in
intervention scenario as presented in Figure 5.

Figure 5. Results of PSA plotted against Base-Case ICER

CE Cloud: IMCI+PO versus IMCI PSA

Incremental Cost simulations
2,000,000.00

1,000,000.00

0.00
0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000

-1,000,000.00

-2,000,000.00

-3,000,000.00

-4,000,000.00

-5,000,000.00

-6,000,000.00
Incremental
QALYs
Basecase Joint incremental cost and LYs
PSA Joint incremental cost and LYs

32
ICER was also compared to the willingness to pay threshold (which is considered to be
one times per capita GDP of the country) and results of PSA were presented in the form
of Cost-Effectiveness Acceptability Curve (CEAC) to compare ICER values with
willingness to pay threshold as presented in Figure 6.

Figure 6: Cost Effectiveness Acceptability Curve

CEAC: % Simulations Accepted at Varying Cost Effectiveness

Threshold (INR/QALYs)
1.005

1
PROBABILITY OF COST-EFFECTIVENESS

0.995

0.99

0.985

0.98

0.975

0.97
0 20000 40000 60000 80000 100000 120000 140000
COST-EFFCTIVENESS/WILLINGNESS TO PAY THRESHOLD

9. Discussion (based on decision model and budget impact analysis)

From the evidence brought forth by our systematic review, it seems that pulse oximetry,
used in conjunction with clinical guidelines like the IMCI, is beneficial in screening and
diagnosis of pneumonia in the community. It is important to note here that such
diagnoses have to be coupled with prompt provision of oxygen therapy at the community
level institutions, in order to reap the benefits of a more early and accurate diagnosis. The
deaths averted due to childhood pneumonia when IMCI+PO is used instead of IMCI alone
is 21 and 36 per 1000 patients when the sensitivity is 70% and 85%. When we take a
lifetime horizon this results in a QALY gain of 1356 and 2339 years respectively.

33
The ICER for both sensitivities show a negative value suggesting that PO when added to
the existing IMCI would become a cost saving intervention. The costing and budget
impact analysis showed that the introduction of pulse oximeter along with existing IMNCI
will increase the cost per patient per year by INR 0.36 only. The overall cost of roll-out of
pulse oximeters in PHC’s would amount to INR 90,459,000. The cost of equipping all the
CHC’s with specified number of pulse oximeters amount to INR 55,100,000. The overall
domestic general government health expenditure per capita for India is US$61.4041. For
a three trillion dollar economy which spends 1% of its GDP on healthcare, the
implementation of the IMNCI+PO would cost only 0.003% of its annual budget.

It is important to know how to interpret the information received from oximetry. The
inherent limitation of being a non-invasive technology makes it all the more important
for proper training to be given to frontline health workers. The technical specifications
of Finger Pulse Oximeter for use in Health and Wellness Centres (Table 4) should be
maintained and updated at regular intervals. The relative ease of implementation of a
pulse oximetry-based intervention (even with the assumption of perfect availability)
compared with the elimination of low birth weight or malnutrition makes it an important
candidate for an intervention against pneumonia in resource-poor settings. The decision
tree was able to show that on top of the large reduction in deaths due to pneumonia, the
addition of pulse oximetry to IMCI has the potential to increase the correct treatment of
severe cases. Thus, pulse oximetry appears to be both an effective and cost-effective
option for the government to contemplate implementation of the same in the primary
healthcare institutions.

In the case of IMNCI+PO, the value of ICER was less than the GDP per capita in all
simulations as part of the probabilistic sensitivity analysis. The sensitivity analysis also
showed that majority of the values fell into the right lower quadrant, signifying ICER to
be negative with gain in QALYs and less cost incurred in the intervention scenario. The
implementation of IMNCI+PO imposes only a small increase in the overall budget. With
an overall health system spending of US$61.40 per capita per year 41, this implies a
0.003% increase in budget which appears reasonable, considering the Government of
India’s strong commitment to raise resource allocation to health for achieving universal
health care. Recommending a program or strategy for scale-up merely on grounds of cost-
effectiveness may not be prudent.

34
10. Conclusions and Recommendations
1. IMNCI+PO is a cost saving prognostic tool as compared to IMNCI alone provided
there is supplementary oxygen availability.
2. IMNCI should be the basic prognostic tool for childhood pneumonia but PO is
beneficial in the referral of cases. Pulse oximetry in general may be used to measure
oxygen saturation in cases wherever required.
3. Among outpatients with pneumonia, oxygen saturations <90% were associated with
increased morbidity and mortality. A hospital admission threshold of <92% would
be safer and clinically better justified. All severe cases irrespective of availability of
Pulse oximeter will be referred to a tertiary care facility for expert management.
4. In tertiary care, when SpO2 ≥ 80%, pulse oximetry has high accuracy in estimating
SaO2 and may be used instead of ABG; in patients with SpO2 < 80%, however, the
exact estimation of SaO2 and the evaluation of oxygenation by pulse oximeter is not
a good substitute for ABG analyzer.
5. Pulse oximeter specification may be as mentioned in Table 4.
6. In tertiary care hospitals, especially in ICU’s Parameters multipara monitors which
measures ECG, Respiration, Pulse Rate, Temperature, SPO2, NIBP suitable for adult
and neonates should be used.

11. Limitations and assumptions

There were several limitations to our analysis. One major limitation of the study was a
lack of data on the availability of oxygen support and how it is distributed throughout the
health system. The IMNCI was functional (at least on paper) in most of the states in India.
But the mortality rate of children in the age group 0-5 is different for different states.
There is still no good quality evidence as to why there is a disparity in the death rates due
to pneumonia despite most states opting to implement IMNCI. So, the question of whether
it is the poor implementation of IMNCI or whether IMNCI as a tool is unsatisfactory is not
conclusively proven. The assumptions made for the decision model and the economic
evaluation is given in Table 14.

35
12. References:
1. Rudan I, Boschi-Pinto C, Biloglav Z, Mulholland K, Campbell H. Epidemiology and
etiology of childhood pneumonia. Bull World Health Organ. 2008 May;86:408-
416B.

2. Liu L, Oza S, Hogan D, Perin J, Rudan I, Lawn JE, et al. Global, regional, and national
causes of child mortality in 2000–13, with projections to inform post-2015
priorities: an updated systematic analysis. The Lancet. 2015 Jan
31;385(9966):430–40.

3. Pletz MW, Rohde GG, Welte T, Kolditz M, Ott S. Advances in the prevention,
management, and treatment of community-acquired pneumonia. F1000Research
[Internet]. 2016 Mar 8 [cited 2019 Jul 27];5. Available from:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4786904/

4. Izadnegahdar R, Cohen AL, Klugman KP, Qazi SA. Childhood pneumonia in

developing countries. Lancet Respir Med. 2013 Sep 1;1(7):574–84.

5. Rodrigues CMC. Challenges of Empirical Antibiotic Therapy for Community-

Acquired Pneumonia in Children. Curr Ther Res. 2017 Jan 1;84:e7–11.

6. Lim YW, Steinhoff M, Girosi F, Holtzman D, Campbell H, Boer R, et al. Reducing the
global burden of acute lower respiratory infections in children: the contribution of
new diagnostics. Nature. 2006 Dec 1;444 Suppl 1:9–18.

7. Basnet S, Adhikari RK, Gurung CK. Hypoxemia in children with pneumonia and its
clinical predictors. Indian J Pediatr. 2006 Sep 1;73(9):777–81.

8. Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M, et al.
ROBINS-I: a tool for assessing risk of bias in non-randomised studies of
interventions. BMJ. 2016 Oct 12;355:i4919.

9. Rudan I, O’Brien KL, Nair H, Liu L, Theodoratou E, Qazi S, et al. Epidemiology and
etiology of childhood pneumonia in 2010: estimates of incidence, severe morbidity,
mortality, underlying risk factors and causative pathogens for 192 countries. J Glob
Health. 2013 Jun;3(1):010401.

10. Farooqui H, Jit M, Heymann DL, Zodpey S. Burden of Severe Pneumonia,

Pneumococcal Pneumonia and Pneumonia Deaths in Indian States: Modelling
Based Estimates. PLoS ONE [Internet]. 2015 Jun 18 [cited 2019 Jul 24];10(6).
Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4472804/

11. Pitt C, Roberts B, Checchi F. Treating childhood pneumonia in hard-to-reach areas:

A model-based comparison of mobile clinics and community-based care. BMC
Health Serv Res. 2012 Jan 10;12(1):9.

12. Hazir T, Nisar YB, Abbasi S, Ashraf YP, Khurshid J, Tariq P, et al. Comparison of Oral
Amoxicillin with Placebo for the Treatment of World Health Organization–Defined
Nonsevere Pneumonia in Children Aged 2–59 Months: A Multicenter, Double-Blind,

36
 Randomized, Placebo-Controlled Trial in Pakistan. Clin Infect Dis. 2011 Feb
1;52(3):293–300.

13. Sinha A, Levine O, Knoll MD, Muhib F, Lieu TA. Cost-effectiveness of pneumococcal
conjugate vaccination in the prevention of child mortality: an international
economic analysis. The Lancet. 2007 Feb 3;369(9559):389–96.

14. Källander K, Hildenwall H, Waiswa P, Galiwango E, Peterson S, Pariyo G. Delayed

care seeking for fatal pneumonia in children aged under five years in Uganda: a
case-series study. Bull World Health Organ. 2008 May;86:332–8.

15. le Roux DM, Myer L, Nicol MP, Zar HJ. Incidence and severity of childhood
pneumonia in the first year of life in a South African birth cohort: the Drakenstein
Child Health Study. Lancet Glob Health. 2015 Feb 1;3(2):e95–103.

16. Foran M, Ahn R, Novik J, Tyer-Viola L, Chilufya K, Katamba K, et al. Prevalence of

undiagnosed hypoxemia in adults and children in an under-resourced district
hospital in Zambia. Int J Emerg Med. 2010 Nov 11;3(4):351–6.

17. McCollum ED, Ginsburg AS. Outpatient Management of Children With World Health
Organization Chest Indrawing Pneumonia: Implementation Risks and Proposed
Solutions. Clin Infect Dis Off Publ Infect Dis Soc Am. 2017 Oct 16;65(9):1560–4.

18. McCollum ED, King C, Deula R, Zadutsa B, Mankhambo L, Nambiar B, et al. Pulse
oximetry for children with pneumonia treated as outpatients in rural Malawi. Bull
World Health Organ. 2016 Dec 1;94(12):893–902.

19. Straus WL, Qazi SA, Kundi Z, Nomani NK, Schwartz B. Antimicrobial resistance and
clinical effectiveness of co-trimoxazole versus amoxycillin for pneumonia among
children in Pakistan: randomised controlled trial. Pakistan Co-trimoxazole Study
Group. Lancet Lond Engl. 1998 Jul 25;352(9124):270–4.

20. Mathur S, Fuchs A, Bielicki J, Van Den Anker J, Sharland M. Antibiotic use for
community-acquired pneumonia in neonates and children: WHO evidence review.
Paediatr Int Child Health. 2018;38(sup1):S66–75.

21. Bassat Q, Lanaspa M, Machevo S, O’Callaghan-Gordo C, Madrid L, Nhampossa T, et

al. Hypoxaemia in Mozambican children <5 years of age admitted to hospital with
clinical severe pneumonia: clinical features and performance of predictor models.
Trop Med Int Health. 2016 Sep;21(9):1147–56.

22. Fu LY, Ruthazer R, Wilson I, Patel A, Fox LM, Tuan TA, et al. Brief hospitalization
and pulse oximetry for predicting amoxicillin treatment failure in children with
severe pneumonia. Pediatrics. 2006 Dec;118(6):e1822-1830.

23. Kelly JM, Osamba B, Garg RM, Hamel MJ, Lewis JJ, Rowe SY, et al. Community Health
Worker Performance in the Management of Multiple Childhood Illnesses: Siaya
District, Kenya, 1997–2001. Am J Public Health. 2001 Oct 1;91(10):1617–24.

24. Madico G, Gilman RH, Jabra A, Rojas L, Hernández H, Fukuda J, et al. The role of
pulse oximetry. Its use as an indicator of severe respiratory disease in Peruvian

37
 children living at sea level. Respiratory Group in Peru. Arch Pediatr Adolesc Med.
1995 Nov;149(11):1259–63.

25. Alwadhi V, Dewan P, Malhotra RK, Shah D, Gupta P. Tachypnea and Other Danger
Signs vs Pulse Oximetry for Prediction of Hypoxia in Severe Pneumonia/Very
Severe Disease. Indian Pediatr. 2017 Sep 15;54(9):729–34.

26. Chinbuah MA, Abbey M, Kager PA, Gyapong M, Nonvignon J, Ashitey P, et al.
Assessment of the adherence of community health workers to dosing and referral
guidelines for the management of fever in children under 5 years: a study in
Dangme West District, Ghana. Int Health. 2013 Jun 1;5(2):148–56.

27. Bewick T, Greenwood S, Lim WS. What is the role of pulse oximetry in the
assessment of patients with community-acquired pneumonia in primary care?
Prim Care Respir J. 2010 Aug 2;19(4):378–82.

28. Majumdar SR, Eurich DT, Gamble J-M, Senthilselvan A, Marrie TJ. Oxygen
Saturations Less than 92% Are Associated with Major Adverse Events in
Outpatients with Pneumonia: A Population-Based Cohort Study. Clin Infect Dis.
2011 Feb 1;52(3):325–31.

29. Razi EM, Akbari H. A Comparison of Arterial Oxygen Saturation Measured Both by
Pulse Oximeter and Arterial Blood Gas Analyzer in Hypoxemic and Non-hypoxemic
Pulmonary Diseases. In 2006.

30. Our product [Internet]. Lifebox. [cited 2019 Jul 28]. Available from:
https://www.lifebox.org/about-lifebox/our-product/

31. Product Details [Internet]. [cited 2019 Jul 28]. Available from:
https://supply.unicef.org/unicef_b2c/app/displayApp/(cpgsize=5&layout=7.0-
12_1_66_69_115_2&uiarea=2&carea=4FB33BED49B94843E10000009E710FC1&c
pgnum=1&citem=4FB33BED49B94843E10000009E710FC14F74A09668E12ACDE
10000009E71143B)/.do?rf=y

32. Troeger C, Forouzanfar M, Rao PC, Khalil I, Brown A, Swartz S, et al. Estimates of
the global, regional, and national morbidity, mortality, and aetiologies of lower
respiratory tract infections in 195 countries: a systematic analysis for the Global
Burden of Disease Study 2015. Lancet Infect Dis. 2017 Nov;17(11):1133–61.

33. Floyd J, Wu L, Hay Burgess D, Izadnegahdar R, Mukanga D, Ghani AC. Evaluating the
impact of pulse oximetry on childhood pneumonia mortality in resource-poor
settings. Nature. 2015 Dec;528(7580):S53–9.

34. Ginsburg AS, Tawiah Agyemang C, Ambler G, Delarosa J, Brunette W, Levari S, et al.
mPneumonia, an Innovation for Diagnosing and Treating Childhood Pneumonia in
Low-Resource Settings: A Feasibility, Usability and Acceptability Study in Ghana.
PloS One. 2016;11(10):e0165201.

35. Spence H, Baker K, Wharton-Smith A, Mucunguzi A, Matata L, Habte T, et al.

Childhood pneumonia diagnostics: community health workers’ and national

38
 stakeholders’ differing perspectives of new and existing aids. Glob Health Action.
2017;10(1):1290340.

36. Technical Specifications | National Health Systems Resource Centre, MoHFW,

Government of India [Internet]. [cited 2019 Aug 21]. Available from:
http://nhsrcindia.org/category-detail/technical-specifications/ODgz

37. Prinja S, Bahuguna P, Mohan P, Mazumder S, Taneja S, Bhandari N, et al. Cost

Effectiveness of Implementing Integrated Management of Neonatal and Childhood
Illnesses Program in District Faridabad, India. PloS One. 2016;11(1):e0145043.

38. State/UT-wise Functioning of Primary Health Centres as on 31st March 2017

[Internet]. Open Government Data(OGD) Community. [cited 2019 Aug 17].
Available from: https://community.data.gov.in/state-ut-wise-functioning-of-
primary-health-centres-as-on-31st-march-2017/

39. Functioning Community Health Centres (CHCs) as on 31.03.2016 [Internet]. Open

Government Data(OGD) Community. [cited 2019 Aug 20]. Available from:
https://community.data.gov.in/functioning-community-health-centres-chcs-as-on-
31-03-2016/

40. Maya C. Way to better lung health. The Hindu [Internet]. 2017 May 3 [cited 2019
Aug 20]; Available from: https://www.thehindu.com/news/national/kerala/way-
to-better-lung-health/article18379283.ece

41. Domestic general government health expenditure per capita, PPP (current
international $) | Data [Internet]. [cited 2019 Aug 19]. Available from:
https://data.worldbank.org/indicator/SH.XPD.GHED.PP.CD?locations=IN

42. Tripathy BS, Specialist D, Estate L. Or UNICEF C4D Specialist, State UNICEF offices.
:24.

39
 13. Addendum
Summary of findings
Table 6: Evidence Summary
No. Title/ Authors/ Journal/ Type of study and Overall outcome Risk of Remarks
Year/ Country number of bias
participants rating -
overall
Evaluating the impact of Modelling study The cost-effectiveness of implementing pulse oximetry in India Moderate Favouring the
pulse oximetry on childhood with PO1 and PO2 are US$11.84 and US$9.56 per DALY, use of pulse
pneumonia mortality in Continuous-time respectively. oximetry
resource-poor settings. deterministic
1
compartmental model.
Floyd J et al.
Nature.
2015 Dec 3
The Role of Pulse Oximetry: Comparative study. Children with pneumonic and non-pneumonic ALRI (59%, Moderate Favouring the
Its Use as an Indicator of 160/269) had a mean (+/- SD) arterial oxygen saturation use of pulse
Severe Respiratory Disease 162 well children significantly lower than well children (93.8% +/- 3.5% vs 98.7% oximetry
in Peruvian Children Living 269 children with ARI +/- 1.51%; P < .01).
at Sea Level
Pulse oximetry misclassified notably fewer well children than did
2 Madico et al. the WHO algorithm (4% vs 35%).
Archives of Pediatrics and
Adolescent Medicine Pulse oximetry and the WHO algorithm together (SATWHO)
1995 detected 99% and 87% of pneumonic ALRI and radiologic
Peru pneumonias, respectively, and both methods detected 94% of all
cases of pneumonic and non-pneumonic ALRI diagnosed
clinically.
Emergency triage Validation study Hypoxemia was strongly associated with inpatient mortality (age- Moderate Favouring the
assessment for hypoxaemia adjusted risk ratio: 4.5; 95% confidence interval, CI: 3.8-5.3). use of pulse
in neonates and young Hypoxemia was found oximetry
3
children in a Kenyan in 977 of 15289 5-15% of the children who had hypoxaemia on admission were
hospital: an observational (6.4%) of all missed, and 18% of the children were incorrectly identified as
study. admissions hypoxaemic.

40
 Mwaniki MK et al. Clinical signs are poor predictors of hypoxemia, and using pulse
Bull World Health Organ. oximetry in resource-poor health facilities to target oxygen
2009 Apr therapy is likely to save costs.
Kenya
Hypoxaemia in Mozambican Hospital-based survey The prevalence of hypoxemia on admission was 27.9%, and Moderate Favouring the
children <5 years of age 19.8% of these children died (OR compared with non-hypoxaemic use of pulse
admitted to hospital with 825 children children 3.22, 95% CI 1.98-5.21, P < 0.001). oximetry
clinical severe pneumonia:
clinical features and None of the models performed well when tested in different case
performance of predictor scenarios of oxygen availability through mathematical modelling,
4
models. with over 50% of hypoxaemic children not receiving oxygen even
Bassat Q et al. in favourable case scenarios.
Trop Med Int Health.
2016 Sep 21 Clinical signs alone or in combination are not suitable to diagnose
Mozambique hypoxaemia. The use of pulse oximeters should be strongly
encouraged.
Prevalence of undiagnosed Cross-sectional Oximetry data changed clinical Moderate Favouring the
hypoxemia in adults and analysis management in all observed cases of hypoxemia and several cases use of pulse
children in an under- of normoxemia, leading to application of supplemental oxygen, oximetry
resourced district hospital in 192 patients initiation of further diagnostic testing, prolongation of inpatient
Zambia. stay, or expedited discharge home.
5
Foran M et al.
Int J Emerg Med.
2010 Nov 11
Zambia
Pulse oximetry for children Survey-based The availability of oximetry appeared to have increased the Moderate Favouring the
with pneumonia treated as assessment referral rate for severely hypoxaemic children without chest use of pulse
outpatients in rural Malawi. indrawing or danger signs from 0% to 27.2% (P < 0.001). In the oximetry
N= 14092 absence of oximetry, if the relevant World Health Organization
6 McCollum ED et al. (WHO) guidelines published in 2014 had been applied, 390/568
Organ. (68.7%) severely hypoxaemic children at study health centres
2016 Dec 1 and 52/84 (61.9%) severely hypoxaemic children seen by
Malawi community health workers would have been considered ineligible
for referral.

41
 The Use of Pulse Oximetry to Retrospective The median pulse oximetry reading of children with radiographic Moderate Do not favour
Exclude Pneumonia in comparison study. pneumonia was 97% (interquartile range 95th- 98th percentile) the use of pulse
Children In our study compared with 98% (interquartile range 96th-99th percentile) in oximetry
population of children the control group.
The American Journal of under the age of 24 Forty-five percent (35 of 78) of children with radiographic
Emergency Medicine months, 803 chest pneumonia showed oxygen saturations of 98% or higher with
2002 Oct radiographs were greater than 10% (8 of 78) displaying oxygen saturations of
USA obtained and 100%.
7 analyzed.
By using logistic regression, pulse oximetry was not found to be a
statistically significant predictive variable for radiographic
pneumonia.

Pulse oximetry could not be used to rule out the presence of

radiographic pneumonia in children less than 2 years of age who
presented with respiratory complaints

42
Table 7: Summary of studies with supportive evidence for pulse oximetry
No. Title/ Authors/ Year/ Country Type of study Overall outcome Risk of Bias - Remarks
and number of Overall
participants

1. Beyond Critical Congenital Heart Comparative The sensitivity and specificity of pulse Moderate Favours
Disease: Newborn Screening Using cross-sectional oximetry screening for non-cardiac diseases expanded use of
Pulse Oximetry for Neonatal Sepsis study. were 42% and 99.9% respectively, and pulse oximetry.
and Respiratory Diseases in a Middle- 100% and 99.7% for CCHD, respectively.
Income Country. N = 5247 Expanded use of pulse oximetry has
immediate implications for low- and middle-
Jawin V et al. income countries contemplating strategies
PLoS One. to reduce neonatal mortality and morbidity.
2015 Sep 11
Malaysia

2. What is the role of pulse oximetry in Prospective cohort SpO₂ ≤ 90% has good specificity but low Moderate Pulse oximetry
the assessment of patients with study sensitivity for adverse outcomes in CAP. complements
community-acquired pneumonia in rather than
primary care? N = 467 replaces clinical
severity scoring.
Bewick T et al.
Prim Care Respir J.
2010 Dec
UK
3. Adoption of paediatric and neonatal Mixed-methods Prior to our intervention, 3.3% of children Moderate Favouring the
pulse oximetry by 12 hospitals in realist evaluation. and 2.5% of neonates had oximetry use of pulse
Nigeria: a mixed-methods realist documented on admission. oximetry
evaluation. Between January
2014 and April In the 18 months of intervention period, all
Graham et al. 2017, 38525 hospitals improved oximetry practices,
BMJ Glob Health. children (38% typically achieving oximetry coverage on
2018 Jun 26 aged ≤28 days) >50% of admitted children after 2-3 months
Nigeria were admitted to and >90% after 6-12 months.
participating
hospitals (23401

43
 pre-training; However, oximetry adoption varied in
15 124 post- different contexts.
training).

4. Brief hospitalization and pulse Post-hoc cohort The inclusion of oximetry data improved the Moderate Favouring the
oximetry for predicting amoxicillin analysis was predictive ability at baseline, 12 hours, and use of pulse
treatment failure in children with nested within a 24 hours. The ability to predict failure after oximetry
severe pneumonia. previously 12 hours of observation with oximetry data
completed, was similar to the predictive ability after 24
Fu LY et al. randomized trial hours without pulse oximetry data.
PEDIATRICS
2006 Dec
9 locations in
8 countries
5. Multi-center study of hypoxemia Prospective, No hospital used pulse oximetry routinely, Low Favouring the
prevalence and quality of oxygen multicenter and only 9 of 40 (22.5%) patients<15 years use of pulse
treatment for hospitalized Malawian observational old with SpO2<90% were treated with oximetry
children. study. oxygen by hospital staff. Study personnel
using WHO criteria for children<5 years old
McCollum ED et al. N= 761 achieved a higher sensitivity (40.0%) and
Trans R Soc Trop Med Hyg. lower specificity (82.7%) than Malawian
2013 May clinicians (sensitivity 25.7%, specificity
Malawi 94.1%).

Although hypoxemia is common, the

absence of routine pulse oximetry results in
most hospitalized hypoxemic Malawian
children not receiving available oxygen
treatment.
6. Can clinical signs predict hypoxaemia Comparative Clinical findings were correlated with Moderate Favouring the
in Papua New Guinean children with cross-sectional different levels of hypoxaemia: <93%, <90% use of pulse
moderate and severe pneumonia? study and <85%. oximetry

Moses Laman et al. N= 77 Cyanosis, head nodding and drowsiness

Annals of Tropical Paediatrics were good predictors of hypoxia but lacked
2005 sensitivity. Decisions to use oxygen based on
PNG these signs would therefore result in a

44
 significant number of children with hypoxia
not receiving oxygen.

Pulse oximetry is the best indicator of

hypoxaemia in children with ALRI and,
although relatively expensive, its use might
be cost-effective in controlling oxygen
requirements.
7. Tachypnea and Other Danger Signs vs Cross-sectional Multiple logistic regression revealed that Moderate Favouring the
Pulse Oximetry for Prediction of study age-specific tachypnea (RR≥70/min for 2-12 use of pulse
Hypoxia in Severe Pneumonia/Very mo, and RR ≥60/min for ≥12 mo), head oximetry
Severe Disease. N= 112 nodding, and inability to drink/breastfeed
were independent predictors for hypoxia
Alwadhi V et al with sensitivity of 70.2%, 50.9% and 75.4%,
Indian Pediatr. respectively; and specificity of 88.9%,
2017 Sep 15 96.4%, and 90.9%, respectively.
India
When all three predictors were used in
conjunction, the sensitivity increased to
91.2% and specificity was 81.8%.

No single clinical sign can perform as well

as pulse oximetry for predicting hypoxia in
children with severe pneumonia.

8. mPneumonia, an Innovation for Qualitative study HCPs and caregivers viewed the pulse Low Favouring the
Diagnosing and Treating Childhood based in ground oximeter and breath counter favorably. use of pulse
Pneumonia in Low-Resource Settings: theory methods Challenges included electricity oximetry
A Feasibility, Usability and requirements for charging and the time
Acceptability Study in Ghana. needed to complete the application. Some
caregivers saw mPneumonia as a sign of
Ginsburg AS et al. modernity, increasing their trust in the care
PLoS One. received. Other caregivers were hesitant or
2016 Oct 27 confused about the new technology. Overall,
Ghana this technology was valued by users and is a
promising innovation for improving quality
of care in frontline health facilities.

45
9. Childhood pneumonia diagnostics: Qualitative CHWs and national stakeholders across the Low Favouring the
community health workers' and methodology, four countries perceived the acute use of fingertip
national stakeholders' differing combination of respiratory infection (ARI) timer and pulse oximetry
perspectives of new and existing aids. pile-sorting fingertip pulse oximeter as highly scalable
activities and and easy for CHWs to use.
Spence H et al. focus group
Glob Health Action. discussions National stakeholders were less receptive to
2017 (FGDs). new technologies. CHWs placed greater
Cambodia, Ethiopia, Uganda and South priority on device acceptability to caregivers
Sudan. and children.

Both groups felt that heavy reliance on

electricity reduced potential scalability and
usability in rural areas. Device simplicity,
affordability and sustainability were
universally valued.
10. Prevalence of hypoxemia in under-five Cross-sectional Of the total number, 42.7% had hypoxemia Moderate Favouring the
children with pneumonia in an study (with pulse oximeter oxygen saturation use of pulse
emergency pediatrics hospital in <90%), out of them 36 (56.25%) were in the oximetry
Sudan. N = 150 age group <2 months. Of the hypoxic
patients, 30 (46.88%) had severe
Salah ET et al. pneumonia, and 7 (10.94) had very severe
Indian J Crit Care Med. pneumonia (P < 0.001).
2015 Apr
Sudan There was a significant association between
the hypoxemia and small age group and
very severe pneumonia. In limited resource
settings pulse oximeter can be used to
correctly identify hypoxemia in under five
children particularly among those
diagnosed clinically as very severe
pneumonia.
11. Prevalence and prediction of Comparative Compared with previous studies of children Moderate Favouring the
hypoxemia in children with cross-sectional living at lower altitudes, the presence of use of pulse
respiratory infections in the Peruvian study tachypnea was relatively nonspecific as a oximetry
Andes. predictor of radiographically determined
N = 423 pneumonia or of hypoxemia, especially in
Reuland DS et al. infants.

46
 J Pediatr.
1991 Dec Radiographic pneumonia was not a sensitive
Peru predictor of hypoxemia or clinically severe
illness. In contrast, the presence of
hypoxemia was a useful predictor of
radiographic pneumonia, with both
sensitivity and specificity of 75% in infants.

We conclude that acute lower respiratory

tract infection in children living at high
altitude is frequently associated with
hypoxemia, and that oxygen should be
administered to children with a diagnosis of
pneumonia in these regions.

Case management algorithms developed in

low-altitude regions may have to be
modified for high-altitude settings. In this
setting, pulse oximetry is a good predictor of
pneumonia. Because pulse oximetry is more
objective and cheaper than radiography, its
role as a clinical and investigative tool
merits further exploration.
12. Accuracy of symptoms and signs in Meta-analysis There was substantial variation in Low Favouring the
predicting hypoxaemia among young sensitivity and specificity between different use of pulse
children with acute respiratory N = 11 diagnostic symptoms and signs as well as across oximetry
infection: a meta-analysis. studies with 5787 studies. Cyanosis, inability to feed, head
patients nodding, respiratory rate > 70/min and
Zhang L et al. unresponsiveness/impaired reusability had
Int J Tuberc Lung Dis. high specificity but low sensitivity. In
2011 Mar contrast, reported rapid breathing and
crepitations in lung auscultation had
relatively high sensitivity but low specificity.
Five models of a combination of symptoms
and signs presented moderate sensitivity
(range 0.60-0.84) and specificity (range
0.63-0.82).

47
 Both single nor combined symptoms and
signs have satisfactory performance in
predicting hypoxaemia among young
children with ARI. Improved access to pulse
oximetry is needed in developing countries.
13. Does pulse oximeter use impact health Systematic Review The evidence is low quality and hypoxaemia Low Favouring the
outcomes? A systematic review definitions varied across studies, but the use of pulse
5 studies included evidence suggests pulse oximeter use with oximetry
Enoch AJ et al. children can reduce mortality rates (when
Arch Dis Child. combined with improved oxygen
AUG 2016 administration) and length of emergency
department stay, increase admission of
children with previously unrecognised
hypoxaemia, and change physicians'
decisions on illness severity, diagnosis and
treatment. Pulse oximeter use generally
increased resource utilisation.

48
Table 8 A: Parameters of Disease Progression

Parameter Value (Range) Sources

Incidence 3.6 million (3.3–3.9 million) episodes Habib Farooqui et al.
of severe pneumonia and 0.35 million Burden of Severe Pneumonia, Pneumococcal
(0.31–0.40 million) all cause Pneumonia and Pneumonia Deaths in Indian
pneumonia deaths occurred in children States: Modelling Based Estimates
younger than 5 years in India.
Rudan, I. et al.
0.075 per child per year Epidemiology and etiology of childhood
pneumonia in 2010: estimates of incidence,
4066541 new severe episodes (severe severe morbidity, mortality, underlying risk
morbidity)/ year factors and causative pathogens for 192
127,960,004 new episodes countries.
(incidence)/ year. J. Glob. Health 3, 010401 (2013).

Proportion severe 5% (2–10%) Pitt, C., Roberts, B. & Checchi, F.

on day 1 Treating childhood pneumonia in hard-to-reach
areas: a model-based comparison of mobile
clinics and community-based care.
BMC Health Serv. Res. 12, 9 (2012).
Mean duration of 3 days (2–4 days) Hazir, T. et al.
non-severe illness Comparison of oral amoxicillin with placebo for
before recovery the treatment of world health organization-
defined nonsevere pneumonia in children aged
2–59 months: a multicenter, double-blind,
randomized, placebo-controlled trial in
pakistan. Clin. Infect. Dis. 52, 293–300 (2011).
Mean duration of 10 days (9–11 days) Hazir, T. et al.
non-severe illness Comparison of oral amoxicillin with placebo for
before progression the treatment of world health organization-
to severe illness defined nonsevere pneumonia in children aged
2–59 months: a multicenter, double-blind,
randomized, placebo-controlled trial in
pakistan. Clin. Infect. Dis. 52, 293–300 (2011).
Mean duration of 4 days (3–5 days) Sinha, A., Levine, O., Knoll, M. D., Muhib, F. &
severe illness Lieu, T. A.
before recovery Cost-effectiveness of pneumococcal conjugate
vaccination in the prevention of child mortality:
an international economic analysis.
Lancet 369, 389–396 (2007).
Mean duration of 7 days (6–8 days) Källander, K. et al.
severe illness Delayed care seeking for fatal pneumonia in
before death children aged under five years in Uganda: a
case-series study.
Bull. World Health Organ. 86, 332–338 (2008).
Proportion 85% viral (75–90%) Le Roux, D. M., Myer, L., Nicol, M. P. & Zar, H. J.
bacterial versus 15% bacterial (10-25%) Incidence and severity of childhood pneumonia
viral (NSV) in the first year of life in a South African birth
cohort: the Drakenstein Child Health Study.
Lancet. Glob. Heal. 3, e95–e103 (2015).

49
Table 8B: Care-seeking and health-care parameters

Mean duration of illness NSV 3 (2–4) days Källander, K. et al.

before care seeking SV 0.75 (0.5–1) days Delayed care seeking for fatal
pneumonia in children aged under five
years in Uganda: a case-series study.
Bull. World Health Organ. 86, 332–338
(2008).
Probability that timely 0.61 ± 10% Walter, N. D. et al.
hospital access Why first-level health workers fail to
follow guidelines for managing severe
disease in children in the Coast Region, the
United Republic of Tanzania.
Bull. World Health Organ. 87, 99–107
(2009).
Probability of community- 0.925 (0.90–0.95) Straus, W. L., Qazi, S. A., Kundi, Z., Nomani,
based treatment curing non- N. K. & Schwartz, B.
severe bacterial case Antimicrobial resistance and clinical
effectiveness of co-trimoxazole versus
amoxycillin for pneumonia among children
in Pakistan: randomised controlled trial.
Pakistan Co-trimoxazole
Study Group.
Lancet 352, 270–274 (1998)
Probability of treatment 0.925 (0.80–0.95) Assumed to be high if oxygen is available
with hospital care curing with lower values representing poorer
case standard of care
Probability of treatment 0.65 (0.6–0.7) Fu, L. Y. et al.
with amoxicillin curing Brief hospitalization and pulse oximetry
severe case if prescription (based on treatment for predicting amoxicillin treatment failure
adhered to failure rates of patients in children with severe pneumonia.
with hypoxia at Pediatrics 118, e1822–e1830 (2006).
baseline)

50
Table 9: Prognostic parameters

Probability of prognostic 1 (0.9–1) Assumed to be high for the purpose of

available this analysis
Sensitivity of IMCI 0.55 (0.5–0.6) Kelly, J. M. et al.
Community health worker performance in
the management of multiple childhood
illnesses: Siaya District, Kenya, 1997–2001.
Am. J. Public Health 91,
1617–1624 (2001).
Sensitivity of PO+IMCI 0.85 (0.8–0.9) Madico, G.
The role of pulse oximetry.
Arch. Pediatr. Adolesc. Med. 149, 1259
(1995).
Specificity of IMCI 0.85 (0.8–0.9) Assumed to be high given low overall
referral rates
Specificity of PO 0.85 (0.8–0.9) Assumed to be similar to IMCI
Adherence to non-severe 0.55 (0.5–0.6) Chinbuah, M. A. et al.
prognosis (IMCI) Assessment of the adherence of community
health workers to dosing and referral
guidelines for the management of fever in
children under 5 years: a study in Dangme
West District, Ghana.
Int. Health 5, 148–156 (2013).

Acácio, S. et al.
Under treatment of pneumonia among
children under 5 years of age in a malaria-
endemic area: population-based surveillance
study conducted in Manhica district- rural,
Mozambique.
Int. J. Infect. Dis. 36, 39–45 (2015).
Adherence to severe 0.65 (0.6–0.7) Chinbuah, M. A. et al.
prognosis (IMCI) Assessment of the adherence of community
health workers to dosing and referral
guidelines for the management of fever in
children under 5 years: a study in Dangme
West District, Ghana.
Int. Health 5, 148–156 (2013).
Adherence to non-severe 0.55 (0.5–0.6) Assumed to be similar to IMCI
prognosis
Adherence to severe 0.85 (0.8–0.9) Assumed to be high for the purpose of this
prognosis analysis

51
Table 10: Cost parameters

Unit Cost Source of information

Amoxicillin INR 47.77 http://janaushadhi.gov.in/old-data/New_MRP_Oct15.pdf
treatment per
child Management Sciences for Health. International Drug
Price Indicator Guide.
http://erc.msh.org/dmpguide/pdf/DrugPriceGuide_201
3_en.pdf (MSF, 2013).
Average hospital INR 82.36 (PHC) World Health Organization. Health Service Delivery Costs
cost per episode INR 116.81 (CHC) and http://www.who.int/choice/cost-
INR 172.80 when the effectiveness/inputs/health_service/en/ (WHO, 2015).
referral is to a tertiary
healthcare facility
Pulse oximeter US$250 Lifebox Foundation.
Lifebox: Saving Lives Through Safer Surgery
http://www.lifebox.org/about-lifebox/our-product/
(Lifebox Foundation, 2015)

INR 2500 Government e-Marketplace Portal

Batteries $2 ($1.5–2.5) Assumed

Uses per set of 840 UNICEF. Supply Catalogue
batteries https://supply.unicef.org/ (UNICEF, 2015).
Lifetime of 2 years UNICEF. Supply Catalogue
device https://supply.unicef.org/ (UNICEF, 2015).
Number of 1 per 1,000 children Assumed
devices needed under 5 (0.8–1.2)

Table 11: Annualization formula

Expected life of equipment Discount Rate Annual Factor

2 0.03 0.5226

Table 12: Cost of treatment37

Per Patient cost of treatment Indexed Figure for 2019

Non-severe pneumonia (home based) 1 OPD Visit + OOP 1117

before recovery
Non-severe pneumonia (home based) 2 (OPD visits + OOP) 2234
before progression to severe illness
Severe pneumonia (hospital based) (1 OPD visit + Per patient 10592
IPD Cost + OOP)

52
Table 13: Cost of training frontline health-worker42

Trainer Honorarium 1000 Per session of 30

Stationery 1000 Per session of 30
DA for Trainee 7500 250/trainee/session
*Infrastructure and logistics 3000 100/trainee/session
TA for trainee 3000 100/trainee/session
Cost of training one person 513.3333
# one session is expected to have 30 frontline health workers
** TA/DA payable to ANM/ ASHA as per state government norms

Table 14: Assumptions made for the decision tree in the study

Assumption
Once a patient is diagnosed with pneumonia, he/ she gets full treatment. Dropout/
poor compliance has not been factored.

Probability of treatment with hospital care curing severe pneumonia is assumed to

be high since we expect oxygen to be available in most FRU’s.

The sensitivity and specificity of diagnosing any severity of pneumonia is the same.
Due to lack of availability of transition probabilities (for e.g.: percentage of
patients going back into non-severe state and not complete cure from severe state
after treatment) some of the intermediate states have been merged.
Proportion of patients with non-severe pneumonia who develop severe
pneumonia, if not given proper treatment

Proportion of patients with non-severe pneumonia who get cured and survive, if
not given proper treatment.
The proportion of children reaching an appropriate health facility may vary
significantly between various states and so having one single parameter for all the
28 states could result in inaccurate estimates

53
Search strings: Pulse oximetry for the diagnosis of pneumonia in children

Pubmed Search

 pulse oxime*[tiab] AND pneumon*[tiab] : 216

 pulse oxime*[tiab] AND pneumon*[tiab] AND child*[tiab] : 89
 (efficacy[ti] OR effectiveness[ti] OR success[ti] OR benefit[ti] OR "Diagnostic Test
Accuracy"[ti]) AND pulse oximet*[ti] : 29
 (sensitivity[ti] OR specificity[ti] ) AND pulse oximet*[ti] :14
 ("positive predictive value"[tiab] OR "negative predictive value"[tiab] ) AND pulse
oximet*[ti] :63
 cost[ti] AND pulse oximet* : 35
 (cost analysis[tiab] OR cost effective*[tiab] OR costing approach[tiab]) AND pulse oximet* :
96
 (Child OR Children OR infant OR infants) AND (Health related quality of life[tiab]) AND
pneumon* : 18
 ((child*[ti] OR infant[ti] OR infants[ti]) NOT adult[ti]) AND pulse oxime* [ti] : 221
Web of Science Search: 140

pulse oxime* AND pneumon* AND child

or
pulse oxime* AND pneumon* AND paedi*
or
pulse oxime* AND pneumon* AND infant

54