Tailored Parenting Plans of Young Adults with Sickle Cell Disease or Sickle Cell Trait

Abstract

Our study purpose was to evaluate the variation and accuracy of tailored parenting plans individually generated as a supplement to reproductive health education on the genetic inheritance of sickle cell disease or sickle cell trait. We present a secondary data analysis of experimental group data from a randomized controlled trial. Participants completed the valid and reliable internet-based Sickle Cell Reproductive Health Knowledge Parenting Intent Questionnaire. We created a computerized algorithm that used participants’ responses to generate tailored parenting plans based on their parenting preferences and partner’s sickle cell status. Thirty-one different parenting plans were generated to meet the variety in the participants’ preferences. The most frequently generated plan was for participants with sickle cell disease who had a partner with hemoglobin AA, who wanted to be a parent, was not likely to be pregnant, and wanted their child to be sickle cell disease-free. More than half of the participants required alteration in their reproductive behavior to achieve their parenting goals. Findings provide insight into the variety and accuracy of computer algorithm-generated parenting plans, which could further guide refinement of the algorithm to produce patient-centered, tailored parenting plans supplemental to internet-based genetic inheritance education.

BACKGROUND

Knowledge of sickle cell disease (SCD) and sickle cell trait (SCT) inheritance is well known to the scientific and health provider community, but very little is known about the plans of young adults with SCD or SCT who desire to become parents. A great deal of progress has been made in SCD management and newer therapies such as stem cell transplantation¹ or gene editing^2,3 have reversed SCD symptoms. However, there is no currently available therapy that eliminates the underlying genetic inheritance, which places individuals with SCD or SCT in a precarious position when they wish to start their families. Individuals with SCD or SCT are still at risk of passing on the gene to their offspring. Whereas investigators have examined the parenting decision-making processes of those living with SCD or SCT⁴ and the efficacy of education programs on sickle cell reproductive knowledge,^5,6 none have investigated the behaviors needed by the SCD and SCT population to put their parenting goals into action. The purpose of this study was to evaluate the variation and accuracy of the tailored parenting plans (PPs) generated for the young adults with SCD or SCT who were randomly assigned to the experimental group of a randomized controlled trial (RCT) focused on reproductive health.

Hemoglobinopathy

In the United States (U.S.), about 100,000 individuals are living with SCD and more than 3.5 million live with SCT.⁷ However, SCD is a hemoglobinopathy of global health concern with an estimated 312,000 babies born with SCD in 2010 alone and an estimated rate of 5,476,407 babies with SCT annually.⁸ Although many individuals with SCT do not have the health complications associated with SCD, they are at risk for having children with SCD. Like individuals with SCD, they may also not be well informed of the genetic inheritance, consequences of passing SCD to their children, and reproductive options available to them.

Parenting Planning

For young adults with SCD or SCT, contemplating the idea of becoming a parent can be a complex, highly involved process particularly if they wish to prevent their children from inheriting SCD. Individuals need to have accurate understanding of genetic inheritance, understand options for becoming parents, and recognize the potential impact of SCD or SCT on their children’s lives and the entire family.^9,10 Individuals living with SCD must consider how to manage their own SCD and that of their children at risk for or living with the disease.⁹ Another consideration is how to determine the appropriate timing of communicating to their children their sickle cell status.¹¹ Timing of pre-parenting information is critically important for young adults as they become sexually active or begin to contemplate starting families of their own.¹¹ Despite these issues, there are very few studies that focus on the PPs or reproductive behaviors of the SCD or SCT population.

We identified two studies that introduced an intervention to educate individuals with SCD or SCT. Using a genetic counseling framework, investigators in one study provided reproductive education to adolescents with SCD.⁵ A pre- and post-test design evaluated the retention of sickle cell genetic inheritance knowledge over a one-year period. However, parenting goals or the development of an actionable PP was not investigated. In the second study, which is our parent study, we explored the possibility of reproductive behavior change using a tailored reproductive educational intervention for young adults with SCD or SCT.^6,12,13 Investigators implemented a web-based, multimedia, interactive intervention providing reproductive decision options to young adults and aided them in developing PPs specific to their goals. The data on the generated PPs from that study ¹³ are the focus of this article.

To date, no published studies have addressed the variety of PPs representing the parenting goals of young adults with SCD or SCT or the behaviors necessary to achieve such goals. The analysis aim was to describe: (1) the variety of tailored PPs generated by a computerized algorithm based on preferences reported by young adults using the Sickle Cell Reproductive Health Knowledge Parenting Intent Questionnaire (SCKnowIQ), and (2) the participants’ perceptions of the accuracy of their system-generated and delivered PPs and video cases consistent with their PP.

METHODS

Original Study: Brief Summary

Details about the original study have been reported elsewhere.^6,13 Briefly, investigators conducted a randomized control trial (RCT) with follow-up over a two-year period. Participants were randomized to the experimental (multimedia, internet-based intervention) group and a control (standard care e-book) group. Both groups received a pre-test at baseline and post-test measures immediately after the intervention (experimental group) or e-book (control group). Participants were compensated $150 over the entire study for their time. The Institutional Review Board at the University of Illinois at Chicago at Illinois approved the study.

We recruited participants from various locations such as sickle cell clinics, community organizations, grocery stores, and through online networks (i.e., Facebook, Craigslist). Data were collected at locations that were convenient for participants.

Participants completed the SCKnowIQ on a touch screen tablet at baseline and immediately after the intervention. The SCKnowIQ measured the RCT outcomes and participants’ preferences about being a parent. Validity and reliability of this assessment instrument have been published elsewhere.¹²

The intervention was an in-depth, online tailored interactive multimedia educational program that provided information on sickle cell inheritance and reproductive options to help young adults with SCD or SCT develop and implement their PPs. As part of the intervention, participants identified, from a series of videos of couples making different reproductive health decisions, the one couple that was most like their situation and decisions they would make. Following completion of the intervention, experimental group participants were presented with their tailored PP, which was generated based on their responses to SCKnowIQ questions and delivered to them immediately on screen and electronically via email. When the PP was generated on screen, each participant was asked to indicate their perception of the accuracy of the PP reflecting their situation and what was inaccurate, if applicable.

Current Study

After receiving exempt approval for use of de-identified data from the Institutional Review Board at the University of Florida, experimental group data were obtained from the original study to conduct a cross-sectional, descriptive secondary data analysis of males and females of reproductive age, 18 to 35 years old who had SCD/SCT. The experimental group included 114 participants (SCD, n = 67; SCT, n = 47). Table 1 represents the sample’s socio-demographic characteristics.

Table 1.

Sample Socio-Demographics

	Frequency	%
Sickle Cell Status
Sickle cell disease	67	59
Sickle cell trait	47	41
Sex
Female	71	62
Male	43	38
Race
Black	51	45
African American	55	48
Multi-racial	4	4
Other	4	4
Ethnicity
Hispanic or Latino	6	5
Not Hispanic or Latino	108	95
Marital Status
Never married	95	83
Married	13	11
Separated	1	1
Divorced	5	4
Education
High school	40	35
Some college	54	47
4-year college degree	13	11
Graduate degree	7	6
Employment
Part-time	21	18
Full-time	22	19
Student	18	16
Not employed	53	46
Income*
<$10,000	42	39
$10,000 – $29,999	31	28
$30,000 – $49,999	18	17
>$50,000	18	17

^*5 participants did not provide income information.

The data source for the current study was the set of PPs generated for the 114 participants. These PPs were generated by a computer algorithm that tailored the PP for each experimental group participant based on their responses to the following five variables from the SCKnowIQ: 1) participant’s SCD or SCT status, 2) partner’s SCD or SCT status, participant 3) wanting to be a parent or not, 4) indicating likelihood of pregnancy, and 5) indicating the importance for a child to be SCD free. The output document, the PP, showed the participant’s behaviors or “special actions” that were necessary to achieve their overall parenting goals with consideration of their other preferences, such as acceptability of adoption or use of advanced fertility methods. Figure 1 is an example of a PP generated for a male participant with sickle cell trait that represented his reproductive goals. As shown in Figure 1, the participant’s answers to the SCKnowIQ items (conditions within the algorithm) are displayed in statements 2–5 and then examples of actions to achieve the goals are listed specific to the conditions. The actions were consistent with information presented in the CHOICES intervention. The PP was intended to be a tailored, take-home summary of the behavior-relevant content for each participant.

Figure 1:

Sample of Computerized Algorithm Generated Parenting Plan

Analysis

We conducted data analysis using the statistical computing software, R.¹⁴ Using descriptive statistics, we examined the unique characteristics of the PPs and their frequency distribution. We used the statistical procedure of compact classification to create a histogram of the PP distribution and show the categories of PPs based on the above-mentioned variables.

RESULTS

PP Frequencies

For the 114 participants’ responses to the SCKnowIQ, the algorithm generated 31 unique versions of the PP. Of these 31 unique PPs, 21 represented the majority of participants with the occasional PP being the same for one or two participants. Figure 2 shows the full frequency distribution of the 31 PPs that were generated for this sample of participants.

Figure 2:

Histogram showing the distribution of the 31 unique parenting plans

Note. y axis labels represent the responses to the following questions from left to right: 1. SC status (0=SCD; 1=SCT)

2. Partner SC status (0=SCD; 1=SCT; 2=Normal; 3=No partner; 4=Unknown)

3. Want to be parent (0=No; 1=Yes; 2=Not sure)

4. Likely to be pregnant (0=No; 1=Yes)

5. Important child free of SCD (0=No; 1=Yes)

For example, the first bar on the y axis represents a participant with SCD (0), whose partner also has SCD (0), who wanted to be a parent, (1) who was not likely to be pregnant in the next 6 months (0), and who wanted their child to be free of SCD (1).

For participants with SCD the 50 most frequent PPs were for the:

• 25 participants who had a partner with hemoglobin AA, which means the partner did not inherit the sickle cell disease or trait, who wanted to be parents, pregnancy was unlikely, and shared it was important to have a child free of SCD.

• 9 participants who had a partner with hemoglobin AA, wanted to be a parent, were likely to be pregnant, and wanted a child free of SCD; and

• 16 participants who had no partner, wanted to be parents, were not likely to be pregnant, and wanted their child free of SCD.

For participants with SCT the 23 most frequent PPs were for:

• 8 participants who reported a partner with hemoglobin AA, wanted to be a parent, were not likely to be pregnant, and wanted their child to be SCD free;

• 6 participants who had a partner with SCT, wanted to be a parent, were not likely to be pregnant, and wanted their child to SCD free; and

• 9 participants who had no partner, wanted to be a parent, were not likely to be pregnant, and wanted their child to be SCD free.

A common feature of the most frequent PPs for participants with SCD or SCT was wanting a child SCD free.

Special Actions to Implement PPs

Table 2 shows whether special action was needed to implement their parenting goals and if the generated parenting plan was accurate as determined by whether a subject was likely to be a parent, considered having a child SCD free important, and chance of a child having SCD. Not all the PPs specified that special action was needed to achieve the participant’s parenting goals and preferences. There were 59 participants not requiring special actions to implement their PPs, they either indicated not likely to have a child, it was not important that their child be free of SCD, or their child would have 0% chance of SCD given the parents’ sickle cell status. Fifty-five participants, however, required implementation of reproductive health behaviors listed in their PPs to achieve their goals regarding becoming parents. For the participants who were likely to have a child, it was important that the child be SCD free. Of these 55 participants, 42 required reproductive health behaviors concordant with their goals because they did not know their partner’s sickle cell status. The other 13 participants were at risk (25% to 100%) of having a child with SCD and having a SCD free child was important to them. Due to a logic flaw in the algorithm generating the PPs for these 13 at-risk participants, five participants were not informed of the special action needed to implement their PPs.

Table 2.

Distribution of Parenting Plans to Achieve Preferences with Perceived Accuracy of Parenting Plan

Likely parent	Child SCD free important	Child SCD chance	Number of Parenting Plans Generated	Special Action		Plan Accurate
				Yes	No	Yes	No	Not sure	N/A
No			8	0	8	7	0	0	1
Yes	No		4	0	4	4	0	0	0
	Yes	0%	47	0	47	41	3	0	3
		25%	10	10	0	10	0	0	0
		50%	1	1	0	0	1	0	0
		100%	2	2	0	2	0	0	0
		Unknown	42	42a	0	32	5	3	2
Total			114	50	59	96	9	3	6

^a5 of these 42 participants received PP generated in error because of a logic flaw in algorithm that generated parenting plans.

Accuracy of PPs

As shown in Table 2, 12 participants reported that the PPs they received were not accurate and did not reflect their desires or situations. One participant commented that the PPs were not sufficiently sensitive to Lesbian, Gay, Bisexual, Transgender, Queer plus (LGBTQ+) family situations. All 12 participants had indicated they were not at all likely to have an abortion to prevent their child from having SCD, and several indicated an objection to abortion as a suggested method to achieve their parenting goals. Table 3 lists an example of participants comments about reasons the PP were perceived as inaccurate.

Table 3.

Participants’ comments about reasons PP were perceived as inaccurate.

Comment	Response to abortion question*
Again, because I am not involved in a heterosexual relationship, this parenting plan does not accurately reflect my choices.	Not very likely
Because I still want to be a father of my own	Not at all likely
Considering ending the pregnancy is not accurate. Neither is ART**	Not at all likely
I don’t want to have any children right now but if I did have children I would want it to be natural and if I did have an abortion it would be because of my health not to prevent myself from having a baby with sickle cell trait or disease	Not at all likely
I have a 50 percent chance to have a child with sickle cell disease	Not at all likely
I just want a healthy child, I don’t care if he/she has the disease or trait	Not at all likely
I would like to have a child that is sickle cell free. Not sure how "important" it is. Would not go through the special testing or fertilization to make sure.	Not at all likely
I would never consider ending a pregnancy.	Not at all likely
I would not end the pregnancy unless my health was in danger.	Not at all likely
I’m uneasy about ending a pregnancy and concerned about the Risk of certain medical procedures to rid a fetus of sickle cell. It may be more risky than dealing with a child with the disease.	Not at all likely
Seeking artificial insemination with eggs or sperm without the sickle cell gene and considering ending the pregnancy.	Not at all likely

^*Abortion question: If you get pregnant in the next 6 months, how likely is it that you will have an abortion because of sickle cell DISEASE in the baby?

^**ART = advanced reproductive technologies.

Response to Videos

Participants were grouped into categories based on their likelihood to become a parent, the importance of the child being SCD free, and the chance of the child having SCD (calculated from the participant’s and their partner’s sickle cell status). Overall, 44 participants identified with scenario 0, the natural pregnancy scenario, and 26 of them had 0% risk of their child having SCD. For the 15 participants identifying with scenarios 1–4, four participants had 0% risk and 11 participants had 25% or unknown risk of their child having SCD. For the other 55 participants, 37 did not identify with any of the scenarios, 15 did not know if they identified with the scenarios, and 3 did not answer the question about the scenarios. Table 4 shows the distribution of participants that responded to the video question shown in Figure 3.

Table 4.

Distribution of Participants’ Responses to Video Scenario Question

Likely parent	Child SCD free important	Child SCD chance	n	Answer to Video Question#
				Scenario 0	Scenario 1–4	None	Don’t know	No answer
No	Yes	0%	4	3	0	1	0	0
		25%	2	1	1	0	0	0
		50%	1	0	0	1	0	0
		Unknown	1	0	1	0	0	0
Yes	No	0%	2	1	0	1	0	0
		Unknown	2	0	0	0	2	0
	Yes	0%	47	22	4	15	5	1
		25%	10	1	5	3	0	1
		50%	1	1a	0	0	0	0
		100%	2	1	0	0	0	1
		Unknown	42	14b	4	16	8	0
Total			114	44	15	37	15	3

Key: Scenario 0= couple one has sickle cell trait, one has normal hemoglobin, and they decided to have pregnancies naturally; 1= couple both have sickle cell trait and they decided to break up; 2=both of couple have sickle cell trait, are using birth control and they will consider adoption, foster care, and other reproductive options; 3= both of couple have sickle cell disease, and they used in vitro fertilization with donor sperm to have their children; 4= one of couple has sickle cell trait, one has sickle cell C trait, and they will consider testing their unborn child to either continue or terminate the pregnancy.

^aConsidered the generated parenting plan inaccurate.

^b3 of 14 considered the generated parenting plan inaccurate.

Figure 3:

Screenshot of video scenario options from CHOICES

DISCUSSION

This study is the first to describe the variety and accuracy of tailored PPs for young adults with SCD or SCT, their perceptions of the accuracy of their tailored PPs, and their responses to which video is closest to their PP. The PPs generated by the computerized algorithm followed an internal logic based on participants’ self-reported responses and was tailored to each participant. We found that the complexity of the participants’ situations and preferences resulted in 31 different PPs to meet the needs of 114 participants. Generally, the participants judged the PPs accurate to their situations. Participants who reported inaccuracies with their PPs were mostly concerned about the suggestion of abortion as one of the options available to them. With deeper analysis of these cases, we noted that they had indicated in the SCKnowIQ that they would not be likely to have an abortion if it was determined their child had SCD. Yet, our algorithm did not include that item. Additionally, we discovered a logic error in the PPs generation algorithm that affected five participants.

Although our results are the first to demonstrate the need for a variety of tailored PPs, previous research addressed some of the aspects affecting the planning process.^4,10 Most of our participants with SCD were not likely to plan a pregnancy despite their desire for parenthood. This finding is consistent with a focus group study indicating that sickle cell status and the perception of its severity play an important role in reproductive planning.¹⁵ Previous researchers described how being a parent to a child with SCD reduced the likelihood to have future pregnancies.^4,6,10 These researchers also emphasized the important role of partners in decisions related to pregnancy. Therefore, partners’ sickle cell status influenced the reproductive plans of young adults with SCD or SCT.

We also found that although several participants indicated the importance of having a child free of SCD and were at some risk for having children with SCD, they identified with the natural pregnancy video scenario. Of the 18 participants in this category, three participants reported inaccuracies with their generated PPs. It is hard to determine if those three individuals were also concerned with the idea of abortion as an option or if there was some other reason for their perceived inaccuracy. Furthermore, there is the possibility that the participants may not have been adequately engaged with the intervention and chose a response to the question at random. Both situations call for further investigation in future studies.

Findings of the current study indicate that due to complexity of participants’ situations, a large variety of different plans are required to address the young adults’ personal parenting intentions. We cannot, however, conclude that the computer algorithm generated PPs represent all the plans needed by the entire populations of young adults with SCD or SCT. Several studies supported the use of algorithms to deliver health messages.^16–18 However, tailoring algorithms are sometimes at risk for logical errors, which was one of our findings. The logic error in the PP generation algorithm prevented five participants from receiving the behavioral actions needed to implement their PPs. We also discovered insufficient sensitivity for the algorithm to generate PPs relevant to the LGBTQ+ community. Although the original study authors conducted a small pilot study of 10 participants,¹² these issues identified in the larger sample of 114 participants were not identified until our team completed this analysis. We recommend that these issues be corrected before the intervention is used in another study. In future studies, these findings can be used to correct errors and refine the algorithm logic to also include responses about abortion intentions as the system generates PPs.

Limitations

This study is not without limitations. The analysis focused on PPs generated for only the initial intervention session of the original study, which also had two intervention booster sessions,⁶ and it is unknown whether additional insights would be gained from evaluations of those PPs. Another unknown is the effects of inclusion of fewer men, individuals with SCT, and individuals of Hispanic ethnicity on the variability and accuracy of the PPs. Finally, the sample was recruited from only one metropolitan area, which may reduce the applicability of study findings to other geographical regions in the U.S. or around the world.

CONCLUSION

In conclusion, despite the study limitations, the findings provide insights about the variety of PPs needed to represent the preferences and situations of young adults with SCD or SCT. The computer algorithm that generates the PPs requires minor revisions to more accurately present behavioral actions perceived by all participants as relevant to their situations, especially considering their responses to acceptability of pregnancy-related procedures and LGBTQ+ status. Clinicians and other health professionals can use these findings to aid in the delivery of effective reproductive strategies and guidance to individuals and couples with SCD or SCT to further manage their own disease. This type of patient-centered, need-focused care for individuals or couples living with SCD or SCT would empower them with the knowledge required to make the best reproductive decisions for their futures. This innovation holds promise for community-based public health interventions to achieve Healthy People 2030 objectives focused on blood disorders, such as SCD.