Abstract
Background:
Attrition, or loss to follow-up, presents a significant threat to the integrity and validity of longitudinal clinical research. Little is known about predictors of attrition in neonatal clinical research, and no prior studies have examined the how families’ experiences participating in research with their infants influences study compliance.
Objective:
To identify novel factors that were associated with attrition over one year of study follow-up among preterm infants enrolled in the multicenter Prematurity and Respiratory Outcomes Program (PROP) observational study.
Methods:
At discharge, research coordinators estimated likelihood of attrition. Parents completed questionnaires about their experience with the study at discharge and at 1 year corrected age. The primary endpoint was completion of four PROP interviews during the first year. Logistic models were used to evaluate the associations between infant, family, and center-based characteristics and attrition.
Results:
Among 318 children, 283 (89%) met the primary endpoint. In bivariate analyses, less maternal education, more people in the household, public insurance, and site were associated with attrition (p<0.05). Parent survey responses, infant characteristics, and site characteristics were unrelated to attrition. Coordinators’ prediction of attrition was associated with completion of early study interviews; this effect waned over time. In multivariable analyses, less maternal education and more people in the household were the factors most strongly associated with attrition.
Conclusion:
Future neonatal research should evaluate novel strategies to decrease burden associated with study participation and reinforcement of study goals with families who have lower educational level to facilitate participation and decrease attrition bias.
Keywords: Attrition, Clinical Research, Prematurity, Follow-up
INTRODUCTION
Attrition, or loss to follow-up, threatens the integrity and validity of longitudinal clinical research.[1] Only 5% of studies achieve 100% follow-up.[2] The direction of ascertainment bias caused by attrition is uncertain and unpredictable. In a systematic review of neonatal studies that reported 18–24 month corrected age (CA) outcomes, there was a nearly 1% increase in developmental impairment for every 1% decrease in the follow-up rate.[2] Thus, children who were lost were more likely to be unimpaired. Similar studies have predicted that children not evaluated at 2 years would have better cognitive outcomes than those who were seen; however, these predictions cannot be confirmed.[3] In contrast, in three regional cohorts and several small studies, preterm-born children who were difficult to follow (but were eventually seen) and dropouts from research studies or clinical care were more likely to be impaired than those followed with ease. [4–9] Given these conflicting results, efforts to minimize ascertainment bias by reducing attrition are essential.
Prior studies did not consistently identify the same demographic or socioeconomic predictors of attrition in neonatal clinical research, and no prior studies have examined the associations between families’ research participation experiences and study compliance.[4–7] To better understand factors associated with attrition in neonatal clinical research, we performed an ancillary study to the multicenter Prematurity and Respiratory Outcomes Program (PROP).[8] The objective was to identify novel factors measured at the time of hospital discharge that were associated with loss to follow-up at 12 months CA.
METHODS
PROP was a prospective multicenter observational study that enrolled infants with gestational ages <29 weeks at <7 days of life.[8–10] Exclusion criteria were: significant congenital heart disease; upper airway, lung, or chest wall anomalies; significant congenital malformations or syndromes; uncertain viability; or family unlikely to be available for long-term follow-up. Eight hundred thirty-five of 1,597 eligible children were enrolled (52.3%). Reasons families were not approached or consent was not obtained are detailed elsewhere; importantly, families of 123 infants were considered unlikely to be available for follow-up and parents of 128 infants were unavailable for consent conversations.[8] PROP collected extensive in-hospital data, parent questionnaires when the children were 3, 6, 9, and 12 months CA, and infant assessments at 1 year CA. The primary outcome was a measure of respiratory morbidity over the first year of life, incorporating the four parent questionnaires.[10] The PROP clinical centers were academic sites with experience conducting neonatal clinical research and funded by NHLBI as part of a consortium of single-center and multi-center projects. [8]
The PROP Attrition Reduction (PROP-AR) Study was performed as an ancillary to the primary study at 10 PROP sites. Enrolled infants survived until first discharge home. Only first-born twins or higher order multiples were enrolled, and children born to parents who did not speak English or Spanish as a primary language were excluded.
Study Measures
Before the PROP-AR Study began, research coordinator(s) at each site were interviewed about their prior experience in research, interactions with families and clinical staff, institutional clinical research infrastructure, and plans to limit attrition in PROP (Supplemental Material). Interviews were recorded with the permission of the coordinators.
During the week before each participant’s anticipated hospital discharge, the research coordinator completed an on-line survey that asked for the coordinator’s estimation of the likelihood that the family would continue to participate in PROP at 3, 6, 9, and 12 months CA. For example, “What is the probability that this PROP participant will complete the 3 month questionnaire? (Please provide a number between 0 (I don’t think this family will complete the 3 month questionnaire) to 100 (I’m sure this family will follow up at 3 months).” Coordinators were instructed to answer these questions based on their subjective experiences with the participants’ families during the neonatal hospitalization.
The research coordinator then distributed a confidential questionnaire to each participant’s parent or guardian (Figure). The questionnaires were returned to the PROP Data Coordinating Center at the University of Pennsylvania in pre-addressed and stamped envelopes. The 10 Likert style questions in the questionnaire were adapted from two valid and reliable surveys: the Critical Care Family Satisfaction Survey (CCFSS) and the Family Satisfaction with Care in the Intensive Care Unit (FS-ICU24) survey.[11,12] The same questionnaire was administered at a clinic visit or mailed to the home with a self-addressed and stamped return envelope at 12 months CA.
Figure:
PROP-AR Study discharge and 12 month parent questionnaire
Primary Outcome, Analyses and Sample Size
The primary outcome of the PROP-AR Study was completion of all 4 post-discharge questionnaires. Factors potentially predictive of attrition (failure to complete all 4 questionnaires) were divided into 3 categories: coordinator or site-based, parent-based, and infant-based. We performed standard bivariate analyses to compare these characteristics between the group of children for whom all 4 questionnaires were completed (“complete”) and the group of children for whom fewer than 4 questionnaires were completed (“incomplete”). Finally, we used a multivariable logistic regression model to determine which factors were most strongly associated with attrition. Site and all factors that were different between the groups at the p<0.2 level in bivariate testing were included in a full model with “complete participation” as the dependent variable. Then, insignificant factors were removed by backwards elimination until only site and statistically significant factors remained in the final reduced model.
The percent of children with complete participation at each site was used as a site-based measure of attrition. Relationships between this continuous measure and other site characteristics (for example, number of coordinators who worked on the study) were evaluated with standard bivariate tests.
We planned to enroll at least 275 participants in this hypothesis-generating observational study. Analyses were performed with STATA IC/13.0 (College Station, TX) and p<0.05 was considered statistically significant. The PROP Attrition Reduction Study was approved as a modification to the main PROP protocol by the Institutional Review Boards at the University of Pennsylvania and all participating clinical centers; additional consent was not required.
RESULTS
Between July 2012 and June 2014, 330 participants were enrolled in the PROP-AR Study. Of these, 12 were excluded (5 placed in protective custody, 5 died after enrollment, and 2 withdrew from PROP), leaving 318 in the analysis cohort. In this cohort, participation in the PROP follow-up telephone questionnaires was: 304 at 3 months (95.6%), 306 at 6 months (96.2%), 300 at 9 months (94.3%) and 300 at 12 months CA (94.3%). Of those who completed the 12-month questionnaire, 149 (49.7%) did the questionnaire by phone, 148 (49.3%) participated in an in-person visit, and 3 (1%) remained in hospital at 12 months. In general, sites that performed infant pulmonary function testing at 12 months completed the 12-month questionnaire in person. All 4 questionnaires were completed for 283 (89.0%). Baseline medical and socio-demographic characteristics of the children with complete study participation, as compared to those with incomplete participation, are displayed in Table 1. Incomplete participation was significantly associated with lower maternal education, more people living in the household, and public insurance. In bivariate analyses, attrition was significantly associated with site (p=0.012).
Table 1:
Characteristics of Study Participants and their Families
| Characteristica | All study participants (n=318) | Incomplete study participation (n=35) | Complete study participation (n=283) | p- value |
|---|---|---|---|---|
| Female gender | 153 (48.1%) | 16 (45.7%) | 137 (48.4%) | 0.76 |
| Birth weight (grams) | 909+/−221 | 928+/−30 | 906+/−226 | 0.58 |
| Gestational age (weeks) | 26.3+/−1.4 | 26.5+/−1.5 | 26.2+/−1.4 | 0.22 |
| Multiple gestation | 50(15.7%) | 3 (8.6%) | 47(16.6%) | 0.32 |
| Maternal ethnicity Latino | 31 (9.8%) | 5 (14.3%) | 26 (9.2%) | 0.36 |
| Maternal race | 0.32 | |||
| White | 190 (59.8%) | 18(51.4%) | 172 (60.8%) | |
| Mother’s highest level of education completed | 0.024 | |||
| High school | 110(37.8%) | 10(33.3%) | 100 (38.3%) | |
| College degree or higher | 54(18.6%) | 1 (3.3%) | 53(20.3%) | |
| Single parent family | 78/313 (24.9%) | 10/33 (30.3%) | 68/280 (24.3%) | 0.45 |
| Number of people who live in the household (including baby) | 0.010 | |||
| 4–6 people | 173 (54.4%) | 15 (42.9%) | 158 (55.8%) | |
| Private insurance | 86/316(27.2%) | 3/35 (8.6%) | 83/281 (29.5%) | 0.009 |
| Among families who completed the PROP-AR Discharge Questionnaireb | ||||
| Family reports prior experience with NICU | 28/197 (14.2%) | 2/17(11.8%) | 26/180(14.4%) | 0.76 |
| Family reports living “out of town” or not in the city where the hospital is located | 93/194 (47.9%) | 10/17(58.8%) | 83/177 (46.9%) | 0.35 |
| Family reports being familiar with at least one member of the PROP research team | 152/193 (78.8%) | 13/17(76.5%) | 139/176 (79.0%) | 0.81 |
Results are presented as n (%) or mean +/− standard deviation
Results are for families who answered each question on the PROP-AR Discharge questionnaire.
There were no differences between the groups in frequency of common inpatient morbidities (surgical necrotizing enterocolitis, ligation of the ductus arteriosus, culture-proven sepsis, or retinopathy of prematurity). Similarly, post-discharge use of medical technologies (CPAP, oxygen, ventilator, tracheostomy, feeding tubes, and home monitors) was not associated with complete study participation.
In general, the coordinators correctly predicted decreasing compliance over time (Table 2). However, on an individual subject basis, the unadjusted odds that the coordinator’s prediction was correct also appeared to decrease over time. Predicted compliance was highly site dependent (p<0.001). Sites with high mean predicted compliance also tended to have high rates of study completion (r=0.65, p=0.04). All sites provided modest financial incentives to families to participate in calls or study visits. These were generally discussed during the consent process and most coordinators did not think that this practice influenced consent rates or study participation.
Table 2:
Research Coordinator Predictions of Complete Participation in the PROP Study Over Time
| Visit | Mean/SD | Median (IQR) | Odds Ratio for Compliance (95% Cl) | P-value |
|---|---|---|---|---|
| 3 Month | 87 +/− 16 | 95 (85–95) | 1.04(1.02–1.06) | 0.001 |
| 6 Month | 85 +/− 18 | 90 (80–95) | 1.03 (1.01–1.05) | 0.009 |
| 9 Month | 86+/− 17 | 90 (80–95) | 1.02 (1.00–1.04) | 0.049 |
| 12 Month | 84+/− 19 | 90 (75–95) | 1.01 (0.99–1.04) | 0.11 |
A mean of 6.7 (range 3–13) research coordinators were employed at each site over the duration of the PROP study. The total number of study coordinators at each site and the duration of coordinator participation were both unrelated to attrition. Factors related to the research culture at each institution, as reported by the coordinators during the pre-study interviews, were unrelated to attrition.
Parents completed and returned 64% (n=204/318) of the PROP-AR Study discharge questionnaires (site range 38%–100%). 91% of families who returned the discharge questionnaire and 85% of families who did not return the discharge questionnaire had complete participation in the main PROP study (p=0.1). Children for whom the discharge questionnaire was not returned were more likely to be from a single-parent household (34% vs. 20%, p=0.005), and were more likely to have mothers who described themselves as black or African American (49% vs. 32%, p=0.026). Other characteristics between the groups with and without a PROP-AR Study parent discharge questionnaire were similar.
The mean response on the 10 Likert-question parent discharge questionnaires was 4.7 +/−0.5. Thus, parents tended to report a strongly positive experience participating in PROP with their child. Parents were provided with the same questionnaire at 12 months CA, and 191 (60%) of the questionnaires were returned. The mean response was 4.9 +/− 0.2, or “very satisfied.” There was no relationship between parent responses on these two questionnaires and attrition.
In a multivariable logistic model, infant characteristics, parental perspectives on study participation, and site were not significantly associated with attrition. Maternal education less than high school degree (adjusted odds ratio (aOR) 0.38, 95% confidence interval (CI) 0.15–0.94) and living with more than 7 people in the household (aOR 0.54, 95% CI 0.30–0.98) were the factors most strongly associated with attrition.
DISCUSSION
In this study, we explored infant, parent, and site-based factors to ascertain whether these were related to attrition in a longitudinal research study. Lower level of maternal education and higher number of people in the home were the factors most strongly and independently associated with attrition. We suggest that these factors are surrogate measures of socio-economic status. We were unable to identify site characteristics or parental perspectives on the experience of research participation that might also predict attrition. Though research coordinators reported widely different approaches to maintaining contact with families, staffing strategies, and research cultures and there was substantial site variation in study staffing, these characteristics were not associated with study participation.
Attrition has been studied in several selected adult populations and linked with demographic characteristics such as age, cognitive status, and smoking.[13–15] In three randomized trials of behavioral weight loss strategies, attrition increased with lower educational attainment, higher body mass index, lack of insurance, and younger age.[16] In a small study of pediatric burn patients, children who did not take part were significantly different from those that participated and dropouts were more likely to have a mental health diagnosis.[17]
There is little known about factors that are predictive of attrition or strategies to prevent attrition in neonatology. Because research participants enrolled in the newborn period are dependent on caregivers throughout the follow-up period for most studies and clinical programs, predictors of attrition may be related to the participant, the family, or relationships between the two. For example, some have suggested that parents who are unprepared to cope with a child’s illness are less likely to comply with follow-up.[18] A few studies have suggested that lower gestational age, center, female gender, lower maternal educational attainment, maternal race or lack of an intact family are associated with attrition in pediatric and neonatal care.[4–7,19] In a 5-center North American trial, withdrawal or loss to follow-up at 36-months was associated with gestational age, center, and socioeconomic/social support.[4] A 1995 German study reported that children who were multiples were more likely to follow up, while girls and children of mothers with lower educational attainment were more likely to drop out.[5] Similarly, regional Australian studies have reported that children who were difficult to follow at 5 and 8 years were more likely to come from disrupted families and had mothers with less education.[7,9] In the current study, neither in-hospital nor post-discharge medical characteristics of the children were related to attrition but socio-demographic factors were strongly related to study completion. Importantly, socio-demographic factors are also strongly related to both high-risk infants’ developmental outcomes as well as developmental trajectories.[20–23] Thus, attrition in studies such as PROP may be non-random, and attrition among children with differing socio-demographic backgrounds is likely to bias the results of longitudinal studies.[24]
Others have reported that clinical center is one of the factors most strongly predictive of participant loss in longitudinal research.[4] While site was highly significant in bivariate analyses, it was not significant in our multivariate analyses. Despite frequent interaction with site research coordinators, we were unable to explain or reduce the significant variation in rates of PROP-AR Study survey completion across sites. We attempted to elucidate which aspects of local research culture were associated with retention of participants; unfortunately, none of the factors examined were predictive of site-specific attrition rates.
A 2007 systematic review identified multiple narrative reports but no quantitative evaluations of different strategies to prevent attrition in adult populations.[25] This study suggested that successful retention strategies can be categorized into three themes: respect for patients, including their ideas and their time commitment to the research; details related to patient tracking, including procedures for subjects lost to follow-up; and support and training of study personnel.[25] The authors applied these strategies in the adult intensive care unit and reported 1 and 5-year follow-up rates of 91% and 86%, respectively. No similar studies have been published in pediatrics or neonatology. Results of the adult studies cannot necessarily be generalized to children, who depend on parents for continued study participation.
The large cohort, detailed longitudinal follow-up, and multicenter nature of PROP made it ideal for evaluating factors related to successful retention of study participants. While frequent phone contact over one year does not represent usual expectations for neonatal follow-up, many research protocols and clinical scenarios require continued family engagement post-discharge and would benefit from targeted strategies to limit attrition. Frequent contact between the research team and families throughout the hospitalization, prior to discharge, and after discharge may explain at least in part the low rate of attrition in PROP. This approach ought to be incorporated into future neonatal studies that require long-term follow-up.
Our study suggests that infants at high risk for loss to follow-up can be identified before hospital discharge. The relationship between attrition and socio-demographic risk factors suggests some potential solutions. Additional incentives (including non-financial incentives), strategies to make research participation less burdensome, and reinforcement of the goals of the study with families who have lower educational level may facilitate participation. Important next steps in this work will be to develop and test strategies to limit attrition among study participants at high risk for loss to follow-up. Such strategies are critical because differential loss to follow-up may lead to misleading conclusions in research studies.
Low attrition in PROP may have limited the power of the study to detect predictors of attrition. Response rates on the parent questionnaires may have limited our ability to understand the experience of all families with the research study. On the other hand, there was not a significant difference in attrition between families who completed the PROP-AR Study questionnaires and those who did not. Lastly, we do not have data about whether the coordinator who assessed the likelihood of compliance for each family was the same coordinator who obtained the original study consent or the duration of that coordinator’s relationship with the family.
Research coordinators are the key to a successful study team. They build rapport, facilitate communication, and ensure that families enjoy research participation. In this study, coordinators identified families at risk for attrition over at least the first 9 months post-discharge. Therefore, sophisticated parent questionnaires and data collection processes may not be necessary. It may instead be more important for the team to have a dedicated research coordinator who understands the babies and their families. This, paired with a thoughtful and focused approach to maintaining engagement with and supporting children and families who have risk factors for attrition, may be the optimal strategy for limiting attrition. Further research will be essential for developing targeted, evidence-based strategies to reduce attrition bias in neonatal clinical research.
Supplementary Material
ACKNOWLEDGMENTS
This study was supported by PROP grants U01 HL101794, U01 HL101456, U01 HL101798, U01 HL101813, U01 HL101465, U01 HL101800, and RO1 HL105702. We gratefully acknowledge all PROP participants and their families, as well as the PROP staff at the University of Pennsylvania Data Coordinating Center and the Research Coordinators at the PROP clinical sites.
Abbreviations:
- CA
Corrected Age
- PROP
Prematurity and Respiratory Outcomes Program
- PROP-AR
PROP Attrition Reduction Study
APPENDIX
PROP Investigators and Study Sites
Cincinnati Children’s Hospital Medical Center:
Barbara Alexander, RN; Claire Chougnet, PhD; Tari Gratton, PA; James M. Greenberg, MD Cathy Grisby, BSN, CCRCl William Hardie, MD; Alan H. Jobe MD, PhD; Beth Koch, BHS, RRT, RPFT; Karen McDowell, MD; Kelly Thornton BS
Washington University:
Pamela Bates, CRT, RPFT, RPSGT; Claudia Cleveland, RRT; Thomas Ferkol, MD; Aaron Hamvas, MD; Julie Hoffmann, RN; Mark R. Holland, PhD; James Kemp, MD; Philip T. Levy, MD; Laura Linneman, RN; Jayne Sicard-Su, RN; Gina Simpson, RRT, CPFT; Gautam K. Singh, MD; Barbara Warner, MD. T, CPFT Gautam K. Singh, MD Barbara Warner, MD
University of California at San Francisco:
Jeanette M. Asselin, MS, RRT-NPS; Samantha Balan; Philip L. Ballard, MD, PhD; Roberta A. Ballard, MD; Katrina Burson, RN, BSN; Cheryl Chapin; Eric C. Eichenwald, MD; Roberta L. Keller, MD; Amir M. Khan, MD; Leslie Lusk, MD; Dennis W. Nielson, MD, PhD; Elizabeth E. Rogers, MD
Vanderbilt University:
Judy Aschner, MD; Amy B Beller BSN; Candice Fike, MD; Scott Guthrie, MD; Tina Hartert, MD; Nathalie Maitre, MD;Paul Moore, MD; Mark O’ Hunt; Theresa J. Rogers, RN; Odessa L. Settles, RN, MSN, CM; Steven Steele, RN; Marshall Summar, MD; Sharon Wadley, BSN, RN, CLS
Rochester-Buffalo:
Kim Bordeaux, RRT; Pam Brown, RRT; Shannon Castiglione, RN; Carl D’Angio, MD; Julia Epping, AAS, RT; Lisa Flattery-Walsh, RRT; Donna Germuga, RRT, CPFT; Aimee Horan, LPN; Heidie Huyck, BS; Nancy Jenks, RN; Vasanth Kumar, MD; Valerie Lunger, MS; Deanna Maffet, RN; Tom Mariani, PhD; Jane O’Donnell, PNP; Mary Platt, RN; Eileen Popplewell, RRT; Sandra Prentice, CRT; Gloria Pryhuber, MD; Clement Ren, MD; Anne Marie Reynolds, MD, MPH Rita M. Ryan, MD; Michael Sacilowski, MAT; Tanya Scalise, RN, BSN; Kristin Scheible, MD; Timothy Stevens, MD, MPH; Elizabeth Werner, MPH; Jason Zayac, BS
Duke University:
pKim Ciccio, RN; C. Michael Cotten, MD; Kim Fisher, PhD; Jack Sharp, MD; Judith A. Voynow, MD
Indiana University:
Charles Clem, RRT; Stephanie Davis, MD; Susan Gunn, NNP, CCRC; Lauren Jewett, RN, CCRC; Brenda Poindexter, MD, MS
Steering Committee Chair:
Lynn M. Taussig, MD, University of Denver
NHLBI Program Officer:
Carol J. Blaisdell, MD
University of Pennsylvania Data Coordinating Center:
Maria Blanco, BS; Denise Cifelli, MS; Jonas Ellenberg, PhD; Rui Feng, PhD; Howard Panitch, MD; Pamela Shaw, PhD; Ann Tierney, BA, MS
Footnotes
Conflict of Interest Statement:
The authors have no conflicts of interest relevant to this article to disclose.
REFERENCES
- 1.Marcellus L: Are we missing anything? Pursuing research on attrition. [Internet]. Can J Nurs Res 2004. September;36:82–98. [PubMed] [Google Scholar]
- 2.Guillén Ú, DeMauro S, Ma L, Zupancic J, Roberts R, Schmidt B, et al. : Relationship between attrition and neurodevelopmental impairment rates in extremely preterm infants at 18 to 24 months: a systematic review. Arch Pediatr Adolesc Med 2012. February;166:178–184. [DOI] [PubMed] [Google Scholar]
- 3.Castro L, Yolton K, Haberman B, Roberto N, Hansen NI, Ambalavanan N, et al. : Bias in reported neurodevelopmental outcomes among extremely low birth weight survivors. PEDIATRICS 2004. August;114:404–410. [DOI] [PubMed] [Google Scholar]
- 4.Aylward GP, Hatcher RP, Stripp B, Gustafson NF, Leavitt LA: Who goes and who stays: subject loss in a multicenter, longitudinal follow-up study. J Dev Behav Pediatr 1985. February;6:3–8. [PubMed] [Google Scholar]
- 5.Wolke D, Söhne B, Ohrt B, Riegel K: Follow-up of preterm children: important to document dropouts. Lancet 1995. February 18;345:447. [DOI] [PubMed] [Google Scholar]
- 6.Callanan C, Doyle L, Rickards A, Kelly E, Ford G, Davis N: Children followed with difficulty: how do they differ? J Paediatr Child Health 2001. April;37:152–156. [DOI] [PubMed] [Google Scholar]
- 7.Constantine WL, Haynes CW, Spiker D, Kendall-Tackett K, Constantine NA: Recruitment and retention in a clinical trial for low birth weight, premature infants. J Dev Behav Pediatr 1993 Feb;14:1–7. [PubMed] [Google Scholar]
- 8.Pryhuber GS, Maitre NL, Ballard RA, Cifelli D, Davis SD, Ellenberg JH, et al. : Prematurity and respiratory outcomes program (PROP): study protocol of a prospective multicenter study of respiratory outcomes of preterm infants in the United States. BMC Pediatr 2015;15:37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Poindexter BB, Feng R, Schmidt B, Aschner JL, Ballard RA, Hamvas A, et al. : Comparisons and limitations of current definitions of bronchopulmonary dysplasia for the Prematurity and Respiratory Outcomes Program. Ann Am Thorac Soc 2015. December;12:1822–1830. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Keller RL, Feng R, DeMauro SB, Ferkol T, Hardie W, Rogers EE, et al. : Bronchopulmonary dysplasia and perinatal characteristics predict 1-year respiratory outcomes in newborns born at extremely low gestational age: A prospective cohort study. The Journal of Pediatrics 2017. May 17;0 DOI: 10.1016/j.jpeds.2017.04.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wall RJ, Engelberg RA, Downey L, Heyland DK, Curtis JR: Refinement, scoring, and validation of the Family Satisfaction in the Intensive Care Unit (FS-ICU) survey. Crit Care Med 2007. January;35:271–279. [DOI] [PubMed] [Google Scholar]
- 12.Karlsson C, Tisell A, Engström A, Andershed B: Family members’ satisfaction with critical care: a pilot study. Nurs Crit Care 2011. January;16:11–18. [DOI] [PubMed] [Google Scholar]
- 13.Young AF, Powers JR, Bell SL: Attrition in longitudinal studies: who do you lose? [Internet]. Aust N Z J Public Health 2006. August;30:353–361. [DOI] [PubMed] [Google Scholar]
- 14.Chatfield MD, Brayne CE, Matthews FE: A systematic literature review of attrition between waves in longitudinal studies in the elderly shows a consistent pattern of dropout between differing studies. J Clin Epidemiol 2005. January;58:13–19. [DOI] [PubMed] [Google Scholar]
- 15.Gades NM, Jacobson DJ, McGree ME, Lieber MM, Roberts RO, Girman CJ, et al. : Dropout in a longitudinal, cohort study of urologic disease in community men. BMC Med Res Methodol 2006;6:58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Goode RW, Ye L, Sereika SM, Zheng Y, Mattos M, Acharya SD, et al. : Socio-demographic, anthropometric, and psychosocial predictors of attrition across behavioral weight-loss trials. Eat Behav 2016. January;20:27–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.McQuaid D, Barton J, Campbell E: Researchers BEWARE! attrition and nonparticipation at large. J Burn Care Rehabil 2003. July;24:203–207. [DOI] [PubMed] [Google Scholar]
- 18.Wariyar UK, Richmond S: Morbidity and preterm delivery: importance of 100% follow-up. Lancet 1989. February 18;1:387–388. [DOI] [PubMed] [Google Scholar]
- 19.Cooper NA, Lynch MA: Lost to follow up: a study of nonattendance at a general paediatric outpatient clinic. Arch Dis Child 1979. October;54:765–769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Manley BJ, Roberts RS, Doyle LW, Schmidt B, Anderson PJ, Barrington KJ, et al. : Social variables predict gains in cognitive scores across the preschool years in children with birth weights 500 to 1250 grams. The Journal of Pediatrics 2015. April;166:870–6.e1–2. [DOI] [PubMed] [Google Scholar]
- 21.Hackman DA, Farah MJ: Socioeconomic status and the developing brain. Trends Cogn Sci (Regul Ed) 2009. February;13:65–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tong S, Baghurst P, Vimpani G, McMichael A: Socioeconomic position, maternal IQ, home environment, and cognitive development. The Journal of Pediatrics 2007. September;151:284–8– 288.e1. [DOI] [PubMed] [Google Scholar]
- 23.Kilbride HW, Thorstad K, Daily DK: Preschool outcome of less than 801-gram preterm infants compared with full-term siblings. PEDIATRICS 2004. April;113:742–747. [DOI] [PubMed] [Google Scholar]
- 24.Doyle LW, Anderson PJ, Burnett A, Callanan C, McDonald M, Hayes M, et al. : Developmental disability at school age and difficulty obtaining follow-up data. PEDIATRICS 2018. February;141:e20173102. [DOI] [PubMed] [Google Scholar]
- 25.Tansey CM, Matté AL, Needham D, Herridge MS: Review of retention strategies in longitudinal studies and application to follow-up of ICU survivors. Intensive Care Med 2007. December;33:2051–2057. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.