Although skepticism towards infant pain characterized much of the 20th century research and clinical practices [
The major challenge with infant pain assessment is that neonates cannot self-report their subjective experience of pain. Moreover, there is a lack of agreement on the best proxy modality of assessing infant pain, whether it is cortical, biochemical, physiological, or behavioural [
Despite the above-mentioned disputes, cardiophysiological indices of pain, such as heart rate (HR) and HR variability (HRV), are pervasive in the hospital setting [
With the assistance of an academic librarian at the University of Toronto, a systematic search was conducted in Medline, Embase, PsychINFO, and CINAHL in July 2014 for English-language references. Searches were limited to articles published from 1970 to 2014 in order to encompass historical and contemporary articles and reviews. Search terms related to acute pain procedures, cardiovascular measures, and infants (0–3 years of age) were systematically paired (see Supplementary File 1 for Medline search in Supplementary Material available online at
We included prospective observational or descriptive studies of individuals equal to or under 3 years of age undergoing an acutely painful procedure, which was monitored using a cardiovascular measure. Our definition of observational studies included cohort studies in which participants were prospectively identified and followed up during acutely painful procedures using cardiovascular indices, as well as cross-sectional studies that observed an acutely painful procedure using a cardiovascular measure across different gestational or postnatal ages. We also included control group data from pain manipulation studies and prospective randomized or randomized controlled trials (RCTs) that investigated the effectiveness of pain management strategies using cardiovascular measures.
Studies were excluded if they described nonhuman animal models of pain, did not measure an acutely painful event nor include a cardiovascular measure of acute pain, were prospective randomized, RCTs, or pain manipulations that did not include a control group, were review articles, case studies, or conference abstracts, or studies that included participants that differed in age at measurement (i.e., collapsing over one or more months), or gestational age (GA) (i.e., collapsing across at least four months of GA). Of note, most studies that were discarded for collapsing over age of measurement were averaging over age spans within infancy greater than 6 months.
Two authors designed the abstract selection criteria with an initial selection of 500 abstracts (Jordana A. Waxman and Rebecca R. Pillai Riddell). Three authors (Angelina Pinhasov, Jordana A. Waxman, and Paula Tablon) independently read and selected from all the retrieved references and abstracts. Any disagreements between reviewers were resolved through discussion. The percent agreement between the raters ranged from 0.96 to 1.0. Full texts of potentially eligible studies were retrieved (see Figure
Included study flow chart following PRISMA guidelines.
A database was created recording GA at birth, postnatal age at measurement, a description of the cardiovascular results, and any covariates that were included when analyzing whether there were differences in cardiovascular measures following an acutely painful medical procedure. It was important to investigate covariates included in the studies, as there are a number of physiological and behavioural variables known to affect the cardiovascular system [
Due to the fact that a gold-standard quality assessment measure was not available for observational studies [
We aimed to synthesize evidence on the development of cardiovascular responses to acutely painful procedures in preterm and term born infants. For qualitative analysis, group-specific data were first separated by age at measurement and subsequently subdivided by GA at birth, as well as cardiovascular outcome measures (i.e., mean heart rate (HR), HR change, maximum HR, total heart rate variability (total HRV), low frequency heart rate variability (LF HRV), high frequency heart rate variability (HF HRV), and low frequency/high frequency ratio (LF/HF ratio)).
We identified 6994 articles from the electronic searches after removal of duplicates. These articles were then reviewed by title and abstract and were included or excluded based on a priori selection criteria. A total of 180 articles were then reviewed by full-text review, and of these, 41 articles (involving 1552 participants) fulfilled the inclusion criteria [
Table
Study characteristics.
Study |
|
Country | Gestational age | Postnatal age |
Acute pain procedure | Cardiovascular measure | Study design | Quality score |
---|---|---|---|---|---|---|---|---|
Abad et al. [ |
15 | Spain | 37–42 | <4 days | Venipuncture | Mean HR | CS; randomized trial | 17 |
Altun-Köroğlu et al. [ |
25 | Turkey | 37–41 | 4–8 days | Heel stick | Maximum HR | CS; double-blind, placebo-controlled trial | 13 |
Bilgen et al. [ |
34 | Turkey | 37–42 | 1–9 days | Heel stick | HR change (%) | CS; randomized trial | 14 |
Bucher et al. [ |
20 | Switzerland | 37–41 | 4 days | Heel stick | HR change (bpm) | CS; randomized trial | 14 |
Campos [ |
20 | United States | 37–42 | 2 days | Heel stick | Mean HR | CS; randomized trial | 15 |
Cong et al. [ |
28 | United States | 28–32 | <14 days | Heel stick | HR increase, HRV | CS; randomized cross-over trial | 17 |
Cong et al. [ |
14 | United States | 30–32 | <9 days | Heel stick | Mean HR, HRV | CS; randomized cross-over trial | 17 |
Craig et al. [ |
56 | Canada | 25–27, 28–30, 31–33, 34–36, and 37–42 | <8 days | Heel stick | Mean HR | CS; observational | 17 |
de Jesus et al. [ |
41 | Brazil | 37–41 | <2 days | Heel stick | HRV | CS; observational | 15 |
de Oliveira et al. [ |
36 | Brazil | 37–41 | <2 days | Heel stick | Maximum HR, HRV | CS; observational | 14 |
Gormally et al. [ |
21 | Canada | 37–42 | 2 days | Heel stick | Mean HR, HRV | CS; randomized controlled trial | 15 |
Goubet et al. [ |
14 | United States | 28–32 | 4 days, 21 days | Heel stick | HR change | C; observational | 13 |
Gray et al. [ |
15 | United States | 37–42 | <3 days | Heel stick | Mean HR | CS; randomized controlled trial | 17 |
Greenberg [ |
21 | United States | 37–42 | <1 day | Heel stick | Mean HRV | CS; randomized trial | 13 |
Grunau et al. [ |
138 | Canada | ≤28, 29–32, and 38–41 | 4 months | Immunization | Mean HR | C; observational | 16 |
Haouari et al. [ |
15 | England | 37–42 | <6 days | Heel stick | HR change (%) | CS; double-blind, placebo controlled trial | 15 |
Jatana et al. [ |
25 | India | 37–42 | <7 days | Heel stick | HR change (bpm) | CS; randomized trial | 10 |
Johnston et al. [ |
20 | Canada | 32–35 | <10 days | Heel stick | Mean HR | CS; randomized cross-over trial | 15 |
Johnston et al. [ |
89 | Canada | 27, 32 | 4 days, 5 weeks | Heel stick | Mean HR | CS; observational | 15 |
Kostandy et al. [ |
19 | United States | 37–42 | 1 day | Hepatitis B vaccination | Mean HR | CS; randomized controlled trial | 17 |
Leite et al. [ |
29 | Brazil | 37–42 | <7 days | Heel stick | Mean HR | CS; randomized clinical trial | 17 |
Lindh et al. [ |
25 | Sweden | 37–42 | 4-5 days | Heel stick | Mean HR, HRV | CS; observational | 11 |
Lindh et al. [ |
28 | Sweden | 37–42 | 3 days | Venipuncture | Mean HR, HRV | CS; randomized, double-blind trial | 15 |
Lindh et al. [ |
45 | Sweden | 37–42 | 3 months | DPT vaccination | Mean HR, HRV | CS; randomized, double-blind, controlled trial | 14 |
Lucas-Thompson et al. [ |
49 | United States | 28–31, 32–34 | 3–5 days, 3–5 weeks | Heel stick | Mean HR | C; observational | 17 |
Oberlander et al. [ |
23 | Canada | 37–42 | 2-3 days | Heel stick | Mean HR, HRV | CS; observational | 13 |
Oberlander et al. [ |
22 | Canada | 37–42 | 2 months | Heel stick | Mean HR, HRV | C; observational | 13 |
Oberlander et al. [ |
12 | Canada | 24–28 | 27–54 days | Heel stick | Mean HR, HRV | CS; observational | 11 |
Ors et al. [ |
34 | Turkey | 37–42 | <9 days | Heel stick | HR change (%) | CS; randomized trial | 15 |
Owens and Todt [ |
20 | United States | 37–42 | 2 days | Heel stick | Mean HR | CS; observational | 12 |
Sajedi et al. [ |
32 | Iran | 37–42 | <1 day | Intramuscular injection | Mean HR | CS; randomized trial | 13 |
Shibata et al. [ |
47 | Japan | 37–42 | 3-4 days | Heel stick | Mean HR | CS; observational | 14 |
Singh et al. [ |
150 | India | 32–34, 35–37, and 37–42 | <7 days | Heel stick | Mean HR | CS; observational | 8 |
Stevens et al. [ |
40 | Canada | 32–34 | <5 days | Heel stick | Mean HR | CS; descriptive | 15 |
Stevens and Johnston [ |
124 | Canada | 32–34 | ≤5 days | Heel stick | Mean HR, maximum HR, and HRV | CS; observational | 17 |
Taksande et al. [ |
80 | India | 37–42 | <7 days | Venipuncture | Mean HR | CS; observational | 11 |
Upadhyay et al. [ |
41 | India | 37–42 | <15 days | Venipuncture | Mean HR | CS; randomized, placebo-controlled, double-blind trial | 16 |
Uyan et al. [ |
21 | Turkey | 37–42 | <11 days | Heel stick | HR change (%), maximum HR | CS; randomized controlled trial | 14 |
Walden et al. [ |
11 | United States | 24–26 weeks | 21 days | Heel stick | Mean HR, maximum HR | C; quasiexperimental, repeated measures | 18 |
Weissman et al. [ |
29 | Israel | 37–42 | 2-3 days | Heel stick | HR increase (bpm), HRV | CS; randomized trial | 11 |
Weissman et al. [ |
24 | Israel | 37–42 | 4–6 days | Heel stick | Mean HR, HRV | CS; randomized controlled trial | 10 |
Generally speaking, a quarter of the studies were from Canada, a quarter from the United States, and a quarter from Europe, with the remaining studies coming from Asia, the Middle East, and Brazil. The majority of studies were randomized trials and encompassed infants born between 24 and 42 weeks GA that were tested between postnatal day 1 and postnatal month 4. The most common acutely painful procedure that was utilized in the studies was heel stick, and mean HR was the most frequently used cardiovascular measure. In terms of the range of quality scores for the papers, the lowest score was 40% [
Age categorizations were difficult to obtain due to the variability between studies in the age groups they analyzed. Based on the available data, the results will be organized by the following postnatal ages (i.e., age at measurement): 7 postnatal days or less, 1 to 2 postnatal weeks, 3 postnatal weeks, and 1, 2, 3, and 4 postnatal months. Since the majority of data are published on infants within the first 7 postnatal days, tables will only be presented for these studies (see Tables
Description of study covariates included in the cardiovascular analyses.
Study | Covariates |
---|---|
Abad et al. [ |
N/A |
Altun-Köroğlu et al. [ |
N/A |
Bilgen et al. [ |
N/A |
Bucher et al. [ |
Sex, nurse, number of lances needed, baseline heart rate, and activity |
Campos [ |
The number of additional sticks required to obtain the blood sample, the duration of the heel stick, the frequency of crying, and the average HR |
Cong et al. [ |
N/A |
Cong et al. [ |
N/A |
Craig et al. [ |
N/A |
de Jesus et al. [ |
Gestational age, birth weight, sex, mode of delivery, diabetic mothers, breast-fed one hour before puncture, and received oral glucose |
de Oliveira et al. [ |
PIPP score in the period before the heel prick |
Gormally et al. [ |
Preintervention baseline (percentage of time crying in the last two minutes before beginning the interventions) |
Goubet et al. [ |
N/A |
Gray et al. [ |
N/A |
Greenberg [ |
Age, weight, time since last feeding, heel stick and blood collection procedure length, and gestational age |
Grunau et al. [ |
Corrected chronological age |
Haouari et al. [ |
N/A |
Jatana et al. [ |
N/A |
Johnston et al. [ |
Apgar scores at 5 minutes, gestational age at birth, time since last painful procedure, number of painful procedures since admission, or received indomethacin in the past 12 hours |
Johnston et al. [ |
Frequency of invasive procedures, severity of illness, ventilation status, and sex |
Kostandy et al. [ |
N/A |
Leite et al. [ |
N/A |
Lindh et al. [ |
N/A |
Lindh et al. [ |
N/A |
Lindh et al. [ |
N/A |
Lucas-Thompson et al. [ |
Number of prior heel sticks, duration of blood draws, sex, and baseline heart rate |
Oberlander et al. [ |
Breast-fed, SSRI exposure, age at time of acute pain, maternal analgesia, dose of SSRI at delivery, and dose of clonazepam at time of delivery |
Oberlander et al. [ |
Breast-fed, SSRI exposure, age at time of acute pain, maternal analgesia, dose of SSRI at delivery, and dose of clonazepam at time of delivery |
Oberlander et al. [ |
N/A |
Ors et al. [ |
N/A |
Owens and Todt [ |
Sex |
Sajedi et al. [ |
Sex |
Shibata et al. [ |
N/A |
Singh et al. [ |
N/A |
Stevens et al. [ |
N/A |
Stevens and Johnston [ |
N/A |
Taksande et al. [ |
N/A |
Upadhyay et al. [ |
N/A |
Uyan et al. [ |
N/A |
Walden et al. [ |
N/A |
Weissman et al. [ |
N/A |
Weissman et al. [ |
N/A |
Mean and standard deviations for heart rate response to acute pain at less than 7 postnatal days.
Gestational age | Reference | Mean HR (bpm) | SD |
---|---|---|---|
25–27 weeks | Craig et al. [ |
172.38 | 17.22 |
|
|||
28–32 weeks | Cong et al. [ |
165.00 | 14.00 |
Craig et al. [ |
168.20 | 10.50 | |
Craig et al. [ |
155.25 | 21.57 | |
Lucas-Thompson et al. [ |
169.27 | 10.89 | |
|
|||
32–34 weeks | Singh et al. [ |
183.40 | 15.93 |
Stevens and Johnston [ |
162.20 | 15.36 | |
Stevens et al. [ |
154.00 | 13.00 | |
Lucas-Thompson et al. [ |
158.18 | 15.19 | |
|
|||
34–37 weeks | Craig et al. [ |
163.20 | 27.82 |
Singh et al. [ |
165.30 | 16.50 | |
|
|||
37–42 weeks | Abad et al. |
170.00 | N/A |
Craig et al. [ |
145.86 | 19.22 | |
Campos [ |
174.00 | 16.60 | |
Gormally et al |
180.00 | N/A | |
Gray et al. |
123.00 | N/A | |
Kostandy et al. |
155.00 | N/A | |
Leite et al. [ |
172.70 | 21.50 | |
Lindh et al. [ |
134.00 | 19.00 | |
Lindh et al. [ |
144.00 | 20.00 | |
Oberlander et al. |
168.00 | N/A | |
Shibata et al. |
170.00 | N/A |
Mean and standard deviations for heart rate change from baseline in response to acute pain at less than 7 postnatal days.
Gestational age | Reference | HR change | SD |
---|---|---|---|
28–32 weeks | Goubet et al. |
0–15 bpm | N/A |
|
|||
37–42 weeks | Altun-Köroğlu et al. [ |
37.00% | N/A |
Bilgen et al. [ |
19.00% | N/A | |
Bucher et al. |
45 bpm | N/A | |
Gray et al. [ |
36–38 bpm | N/A | |
Haouari et al. [ |
11.40% | 3.0 | |
Jatana et al. [ |
31.48 bpm | 6.66 bpm | |
Ors et al. [ |
19.00% | N/A | |
Owens and Todt [ |
49.00 bpm | 17.5 bpm | |
Sajedi et al. [ |
10.81 | N/A | |
Uyan et al. [ |
38.20% | N/A | |
Weissman et al. [ |
36.50 bpm | 19.50 bpm |
Mean and standard deviations for low frequency heart rate variability in response to acute pain at less than 7 postnatal days.
Gestational age | Reference | Mean LF HRV | SD |
---|---|---|---|
28–32 weeks | Cong et al. [ |
17.62 | 24.55 |
|
|||
37–42 weeks | Gormally et al. |
1.65 | N/A |
Lindh et al. [ |
4.2 | 0.4 | |
Lindh et al. [ |
4.00 | 0.39 | |
Oberlander et al. |
11.0 | N/A | |
Weissman et al. [ |
1.45 | 0.38 |
Mean and standard deviations for high frequency heart rate variability in response to acute pain at less than 7 postnatal days.
Gestational age | Reference | Mean HF HRV | SD |
---|---|---|---|
28–32 weeks | Cong et al. [ |
23.52 | 35.96 |
|
|||
37–42 weeks | de Oliveira et al. [ |
0.44 | 0.69 |
Greenberg |
2.5 | N/A | |
Lindh et al. [ |
3.4 | 0.60 | |
Lindh et al. [ |
3.23 | 0.45 | |
Oberlander et al. |
2.0 | N/A | |
Weissman et al. [ |
0.76 | 0.50 |
Mean and standard deviations for low frequency/high frequency ratio in response to acute pain at less than 7 postnatal days.
Gestational age | Reference | Mean LF/HF ratio | SD |
---|---|---|---|
28–32 weeks | Cong et al. [ |
1.75 | 1.84 |
|
|||
37–42 weeks | Oberlander et al. |
6.00 | N/A |
Weissman et al. [ |
6.1 | 3.2 |
Mean and standard deviations for total heart rate variability in response to acute pain at less than 7 postnatal days.
Gestational age | Reference | Mean total HRV | SD |
---|---|---|---|
32–34 weeks | Stevens and Johnston [ |
4.52 | 2.95 |
|
|||
37–42 weeks | Lindh et al. [ |
4.30 | 0.40 |
Lindh et al. [ |
4.10 | 0.35 |
Mean and standard deviations for maximum heart rate in response to acute pain at less than 7 postnatal days.
Gestational age | Reference | Maximum HR (bpm) | SD |
---|---|---|---|
37–42 weeks | Campos [ |
192.00 | 11.80 |
de Jesus et al. [ |
149.00 | N/A | |
Owens and Todt [ |
179.40 | 13.40 | |
Singh et al. [ |
160.30 | 20.00 | |
Taksande et al. [ |
151.00 | 10.40 | |
Uyan et al. [ |
186.00 | N/A |
No studies investigated cardiovascular responses to acute pain in extremely preterm infants in the first or second postnatal week of life.
No studies investigated cardiovascular responses to acute pain in very preterm, moderate to late preterm, or full term infants in the third postnatal week of life.
No studies investigated cardiovascular responses to acute pain in moderate to late preterm or full term infants in the third postnatal week of life.
No studies investigated cardiovascular responses to acute pain in extremely, very, or moderate to late preterm infants in the second postnatal month of life.
No studies investigated cardiovascular responses to acute medical procedure pain in extremely preterm, very preterm, or moderate to late preterm infants in the third postnatal month of life.
No studies investigated cardiovascular responses to acute pain in moderate to late preterm infants in the third postnatal week of life.
To our knowledge, this is the first systematic review investigating the development of cardiovascular indices of acute pain responding across the first year of life. Large gaps were elucidated in this review and suggest that the development of infant pain responding outside of the first month of life still remains largely unknown. By way of overview, when measuring HR in the first 7 days of life, the variability within each age group on these measures became larger as the infant’s GA increased. Measures of HRV in the first 7 days of life seemed to show less variability within age categories as the child’s GA increased. Data from other postnatal age groups (i.e., 2nd week, 3rd week, 1 month, 2 months, 3 months, and 4 months) were very sparse with patterns generally impossible to discern due to the total absence or presence of only 1 study.
The following paragraphs will discuss key findings and patterns in the results of the systematic review with specific attention to GA at birth, age at measurement, and type of cardiac measurement in response to acutely painful procedures. Limitations and review contributions to the literature, as well as key areas for future research based on the findings, will be highlighted.
Those born at less than 28 weeks GA displayed a blunted HR response to acute pain in the first week of life. At three weeks, as well as one and four postnatal months, mean HR was found to significantly increase during acutely painful procedures, as compared to baseline HR. Mean HR was higher at four postnatal months than during the first postnatal month. This synthesis suggests that mean HR responses to acute pain may stabilize developmentally (i.e., to increase in response to a stressor as in older humans) in extremely preterm infants after the first postnatal week of life. At one postnatal month of life, LF and HF HRV and the LF/HF ratio decreased in response to acute pain.
Past research has noted a blunted pain response in the first week of life based on GA [
In infants born between 28 and less than 32 weeks GA, mean HR was found to significantly increase following an acutely painful procedure from birth to four months of age. Mean HR was found to be higher at four months compared to one to two postnatal weeks of life. As mentioned above with extremely preterm infants, this increase in mean HR is likely linked to an increase in the parasympathetic contribution to HR control [
Although HRV components were only investigated in one study of very preterm infants in the second postnatal week of life, LF and HF HRV were found to increase, while the LF/HF ratio decreased in response to acute pain. Caution should be taken when interpreting these HR and HRV results, as it is based on four studies (quality scores range from 62 to 85%) and a single study (quality score: 85%), respectively. Additionally, only one study included covariates in their analysis.
In infants born at 32 to less than 37 weeks GA, mean HR was found to increase in response to acute pain during the first postnatal week of life; however, the magnitude of responses was variable. Mean HR was found to be stable across the second week of life and increased in response to acute pain. The inconsistencies in mean HR may be due to differences in the acute pain procedure (i.e., heel stick versus venipuncture), the variability in quality of studies (40 to 85%), and the lack of covariates included in the analyses of more than half of the studies (3/5).
Additionally, when total HRV in response to acute pain was examined in the first week of life in those born at 32 to 34 weeks GA, it did not significantly differ from baseline HRV. This may be due to nonlinearity in heartbeats, which is necessary to measure HRV, being less apparent before 35 weeks GA [
During the first four postnatal months of life, full term infants displayed an increase in mean HR in response to acute pain; however, the magnitude of responses was variable. A relative increase in mean HR over the first four postnatal months was noted and may reflect a developmental, relative increase in the parasympathetic contribution to HR control [
During the first postnatal week of life, total, LF, and HF HRV in response to acute pain were found to be inconsistent among full term infants. The LF/HF ratio was the only consistent measure of HRV, and it was found to decrease in response to acute pain across studies. At two postnatal months, LF HRV and the LF/HF ratio decreased in response to acute pain. At three months of age in response to acute pain, total and LF HRV increased. There were no differences in HF HRV when compared to baseline levels at two or three postnatal months.
As mentioned above, the inconsistency within the HRV domains may be explained by the linear statistics utilized by authors [
It is possible that we have omitted relevant studies despite our detailed search strategy, and we specifically excluded non-English language studies. Additionally, group-specific data (i.e., age at measurement and GA) were separated based on available data and natural groupings, which on occasion led to overlap in GA groups.
With regard to analyzing HRV, studies differed on spectrum calculation methods and models of data analysis. Although terminology such as LF and HF bands is common in the field, studies differ on frequency limits of the bands. Other studies utilized linear statistical approaches of comparing means and variance, which has been reported as less sensitive in classifying HRV in infants [
Furthermore, it was difficult to draw conclusions across development and GA groups for cardiovascular responses to acute medical procedure pain, as the majority of studies did not include covariates in their analyses that could impact an infants’ cardiovascular response to acute pain. It is important to keep in mind that the variability in mean HR and HRV components may be due to this lack of control within the studies.
The presence of variability in HR in older preterm infants and full term infants presents an important clinical challenge to gold-standard measures such as the PIPP-R, N-PASS, COMFORT, and Bernese Pain Scale. For example, the PIPP-R has physiological items (i.e., HR) that are numerically scored on a four-point scale reflecting changes in each variable from baseline values [
A lack of control within the studies investigated has been highlighted, with only 13 out of 41 studies including covariates in their analyses. Moreover, the covariates utilized in the studies are divergent, which may have increased the amount of variability noted in the cardiovascular responses to acute pain. Future research in the area of infant pain should address this lack of control by identifying and controlling for factors that may affect an infants’ cardiovascular response to acute pain in their own research. Examining the studies that did use covariates, key covariates that should seriously be considered for inclusion in all cardiac response to pain studies (depending on design) are gestational age, age at measurement (i.e., postnatal age, corrected chronological age), birth weight, time since last feeding, ventilation status, baseline (i.e., prehandling cardiac responding), length of painful procedure, number of painful procedures (e.g., how many draw attempts), illness severity, sex, and respiration rate.
To our knowledge, this is the first systematic review of cardiac responding to acute pain within infancy. Forty-one studies were included to examine different methods of cardiac responses within developmentally sensitive categories. In response to acutely painful procedures, most infants had an increase in mean heart rate; however the magnitude of the increase showed great variability. Research in the area of heart rate variability was inconsistent limiting interpretation of studies using this measure. More attention to covariates and agreement on methodological factors related to cardiac measurement is needed to better understand this physiological response to pain.
Jordana A. Waxman is a trainee member of Pain In Child Health (PICH), a strategic research training initiative of the Canadian Institutes of Health Research.
The authors had no conflict of interests to disclose.
This research was funded by salary and operating and infrastructure awards from the Canada Foundation for Innovation, Canadian Institutes for Health Research, and the Ontario Ministry of Research and Innovation awarded to Rebecca R. Pillai Riddell and awards to Jordana A. Waxman from the Lillian Wright Maternal Child Health Scholarship Program and Ontario Graduate Scholarship.