Texting as a method to collect post-discharge data after major pediatric surgery: lessons learned

Introduction

The first weeks at home after a major pediatric surgery are critical for recovery and long-term prognosis.¹ Despite this, we know little about the pain experienced during the early post-discharge period.² A better understanding of recovery during this time could provide insight into factors associated with the development of adverse outcomes (eg, chronic post-surgical pain) and highlight intervention opportunities.

Regarding remote data collection, it is critical that research teams collect both accurate and complete data. Furthermore, because the time following discharge is stressful to families, data collection should occur with the least amount of burden possible.⁴ While some studies have reported on outcomes such as pain and function post-discharge, no single method stands out as feasible, acceptable, and non-intrusive. In addition, assessment methods, timelines, and outcomes have varied considerably across studies, making it difficult to draw conclusions about methodology.^3,5,6 Currently, it is difficult to know which methods provide the most complete and accurate data, and which are the least burdensome to the family and the research team. The study reported here, details our experience and lessons learned from the use of texting + Qualtrics (an online platform) to collect data from pediatric patients and parents during the first two weeks at home after a major surgery. The primary aim was to examine the method feasibility and acceptability. Secondary aims were to explore patient age and sex as factors that may influence data accuracy, in turn, negatively influencing feasibility. We hypothesized that the texting platform would be feasible and acceptable for patients and parents. No a priori hypotheses were specified for the analyses associated with the secondary aims.

Methods

This secondary analysis used combined data from two prospective studies conducted at Children’s Wisconsin from 7/2/2019–7/16/2021; both approved by the hospital Institutional Review Board. Study #1 (unpublished) included children aged 7–17 undergoing spinal fusion, pectus correction, or growing rod placement; study #2 (Brimeyer et al.,⁷) included patients of any age undergoing pectus repair. Both studies required at least one parent able to read and speak English. Caregivers were required to receive text messages, whereas patients could opt in or out of text messages. Starting on post-discharge day (PDD) 1, patients and parents received a single daily text until PDD 7, and a single text every other day PDD 8–14. The texts contained a link (Qualtrics) to complete questions on pain intensity and oxycodone use. All participants received an exit survey with questions about acceptability and feasibility.

Descriptive statistics were computed using SPSS V.27 (IBM, SPSS Inc, Chicago, IL). Continuous variables are presented as mean (±Standard deviation), or median (interquartile range), and categorical variables are presented as n (%). Categorical data were analyzed with either Fisher’s exact tests or Chi-Square. Acceptability analyses involved coding qualitative data from exit surveys. Data were evaluated for recurring themes by a study team member using inductive thematic analysis and an open coding approach. The relationship between parent and patient reports of high pain intensity and oxycodone use were described using Bland-Altman plots. Significance was set at P < .05.

Results

Feasibility and acceptability

Patients: The combined sample from the two studies consisted of 64 participants ranging in age from 8–21 years (Median 14.0, IQR 12.0, 16.0) with 51.5% male (n = 33). A total of 38 (59.4%) youth (17 males and 21 females) opted into texting. Those who opted out (14.0 years, [11.00, 14.3]) were younger than those who opted in (15.0 years, [13.0, 16.0]) (P = 0.001); groups did not differ on sex, race, or ethnicity. Participants responded to 63.0% (±33.1%) of texts, with 15 participants responding to >80%, and 6 responding to zero texts. Overall, this resulted in 62.9% of the participants’ data available for analysis. Almost half of the participants (n = 18; 47.4%) responded to the exit survey.

Parents: The combined sample consisted of 62 parents, who responded to a mean of 68.5% (±31.1) of texts. Thirty-one parents responded to >80% of all texts, and four parents responded to zero texts. This resulted in 71.4% of parents’ data available for analysis. Forty-five (72.6%) parents responded to the exit survey. Data on patient and parent qualitative responses were primarily positive (Table 1). Parent-patient agreement about pain and oxycodone use appears to be influenced by patient age and sex (Figure 1).

Table 1.

Qualitative assessment of the texting system by patient and parent participants.

Question	Recurring Themesa	Other Comments
Question 1: What did you like the most about the texting system?b	Easy n = 8	“For someone my age, I do believe that the texting system is a good alternative to an email or phone call.” “That the texting system always comes at the same time of day for me to do the assessment.” “That it can be done at any time.” “That we don’t have to go all the way to Milwaukee to answer these questions.” “It was really good for data collection and was easy even when it was hard for me to do other things.” “I remember to do it because it texted me.” “I got to reflect on my pain/medication I was taking every day or two.” “I like that the doctors are checking up on me.”
Question 2: What did you like the least about the texting system?c	Nothing n = 10	“Having to do it for over two weeks.” That there is nowhere to add extra comments.” “At first, I felt like I was constantly answering these texts.” “Frequency”
Question 3: What did you like the most about the texting system?d	Easy n = 20 Convenient n = 14 Quick n = 6 Ability to reply at their own convenience n = 5 Superior to alternative methods (email, call, mail) n = 3	“… if I had already done the survey it told me I already did it.” “It was relevant to my daughter's situation.” “I like the reminders.” “Good to see the progress of my child.” “Kept me aware of the amount of times of medication.” “It was nice to know that was expected of the postop course.” “The multiple choice questions.”
Question 4: What did you like the least about the texting system?e	Nothing n = 22 Wanted an “additional comments” section n = 3 timing of texts n = 3	“The question asked about my child seems more geared for kids 10 and under.” “Repetitive questions.” “Some of the questions did not apply because my child is 15 years old.” “Sliding number scale. It would be easier from a drop-down menu.” “Everything was fine, sometimes I wasn’t sure how to answer a question, another category would have been helpful.” “Once we entered our additional medication, I wish it had saved it in our profile and offered it as an option each time so I didn’t have to keep entering it.” “Inconvenience” “Remembering to do it”

^aNumber reflects the number of participants with a similar response. Each participant could contribute to more than one theme.

^b17 of 18 patients answered this question.

^c17 of 18 patients answered this question.

^d43 of 45 parents answered this question.

^e38 of 45 parents answered this question.

Figure 1.

(A) Bland-Altman plot depicting parent-child agreement on reporting high pain intensity as a function of patients’ sex and age group. Overall, there were 89/158 discrepancies between parent and child reports. The bias plot shows that there was greater agreement between parent reports of pain and male children than between parent reports and female children. This was especially true for younger females, where parents both over- and under-estimated the patients’ reported pain. Black line = mean; Thin dashed lines = ± 1.96 SD. (B) Bland-Altman Plot depicting parent-child agreement on oxycodone use as a function of patients’ sex and age group. Although there is no established limit of agreement, given that reports center on opioid use, even a single discrepancy would be unacceptable. Overall, there were 29/158 discrepancies between parent and child reports. The bias plot shows that there was greater agreement between parent reports and male children than between parent reports and female children. This was especially true for younger females, where parents both over- and under-estimated the patients’ oxycodone use. Black line = mean; Thin dashed lines = ± 1.96 SD.

Discussion

This study examined the feasibility and acceptability of text communication + Qualtrics (an online platform) to gather information on post-surgical pain and opioid use at home following a major pediatric surgery. Both patients and parents found the texting platform acceptable. However, while the method was feasible for parents, it was less so for patients. Parents found the system convenient and reported that texting was their preferred communication method. Several unexpected benefits were reported by patients and parents. One noteworthy example: A patient reported that the texts made them feel cared for by hospital staff.

This study offers a number of lessons learned, which may be helpful to those aiming to collect data during the post-discharge period, or other situations requiring remote data collection. First, older patients were more likely to opt into texting than younger patients. Second, convenience is important. Convenience was the most frequently reported positive characteristic of the texting method. Finally, the reporter matters. While some studies have collected data from both patients and parents, others have collected post-discharge data from only one reporter.^5,6,8–10 This study highlights the importance of collecting data from multiple reporters. The latter point is emphasized by the analyses of parent-patient agreement on post-discharge pain and oxycodone use. Given that the data appear to be influenced by patient age and sex, future studies should be designed to control for these factors.

These findings are limited by the homogeneity of the sample. Feasibility and acceptability for those of different races and ethnicities will be important in future research. This study is also limited by a small sample size. Despite these limitations, this study provides important lessons and ideas which can be used to improve study design, when collecting data remotely and during a stressful time for families.

Funding

This project was supported in part by the National Center for Advancing Translational Sciences, National Institutes of Health, Award Number UL1TR001436. The content is solely the responsibility of the author(s) and does not necessarily represent the official views of the NIH. This project is funded in part by the Research and Education Program Fund, a component of the Advancing a Healthier Wisconsin Endowment at the Medical College of Wisconsin.

Conflicts of interest: None declared.