Parental Status in Treatment Decision Making among Women with Nonmetastatic Breast Cancer

Abstract

Background.

Having dependent children may affect cancer treatment decisions. We sought to describe women’s surgery and chemotherapy decisions in nonmetastatic breast cancer by parental status.

Methods.

We conducted a secondary analysis of the 2015 cross-sectional Share Thoughts on Breast Cancer Study, conducted in 7 Midwestern states in the United States, restricted to women of prime parenting age (aged 20–50 years) who consented to the use of their medical records (N = 225). We examined treatment decisions using data visualization and logistic regression (adjusted for age, stage, family history of breast cancer, income, education, race, health insurance, and partner status).

Results.

Women with dependent children received bilateral mastectomy more often than women without dependent children (adjusted odds ratio 3.09, 95% confidence interval 1.44–6.62).We found no differences in the receipt of chemotherapy by parental status. Women reported more active roles in surgery than in chemotherapy decision making.

Conclusions.

As a likely factor in cancer treatment decisions, parental status should be addressed in clinical practice and research. Future research should assess patients’ sense of ownership in treatment decision making by treatment type.

Understanding what matters to patients is crucial in patient-centered care.¹ Being a parent shapes one’s identity and values, and cancer patients with dependent children face unique challenges.^2-6 When asked about hypothetical treatment scenarios, parents with cancer express preferences for aggressive, life-prolonging treatment, arguing that they need to survive for their children, with fewer parents stating that they prefer treatment that maximizes functional status.³ Regardless of specific choice, parents with cancer typically cite child-centered reasons as the driving motivation behind their treatment decision making.^3,7-9

We aimed to describe treatment decisions among young women with breast cancer, exploring the associations between parental status and treatments received (bilateral mastectomy and chemotherapy). We hypothesized that parents would pursue more aggressive treatment, more often undergoing bilateral mastectomy and receiving chemotherapy.

Methods

Study Population

We used data from the 2015 cross-sectional Share Thoughts on Breast Cancer Study, conducted at 8 sites in 7 Midwestern states in the United States. The study was approved by a central Institutional Review Board at the University of Iowa. Study procedures are described elsewhere.¹⁰ Briefly, each site extracted tumor registry data on eligible patients. Women 18 years or older with a diagnosis of nonmetastatic breast cancer between January 2013 and May 2014 and without a prior history of cancer were eligible. Honest brokers selected random samples of up to 250 eligible patients per site who were invited to complete a mailed survey. Of 1986 women who were invited, 1235 (62.2%) returned the survey. Electronic medical records were used for 852 women (69%) who consented to have these data used in the study. We restricted our analysis to women aged 20 to 50 years as the prime age range for parenting dependent children and to those with available medical record data (N = 225).

Study Measures

The 2 primary outcomes were self-report of surgery type (bilateral mastectomy v. unilateral mastectomy or breast conservation) and receipt of chemotherapy (yes/no). Secondary outcomes included participant report of patients’ and providers’ roles in treatment decision making.

The main exposure was whether participants had dependent children, which was not directly measured in the study. Instead, we operationalized a proxy measure, identifying parents as participants aged 20 to 50 who were either 1) married or living with a partner/significant other and reported a household size (number of people supported by the household income at diagnosis) >2 or 2) widowed, divorced, separated, or never married and reported a household size at diagnosis >1. Nonparents were identified as participants aged 20 to 50 years who were either 1) married or living with a partner/significant other and reported a household size ≤ 2 or 2) widowed, divorced, separated, or never married and reported a household size of 1.We tested the validity of this measure in a sensitivity analysis because some women classified as parents might have cared for older adults.

We included potential demographic and clinical confounders (identified a priori) in our adjusted regression model: age at diagnosis, cancer stage (0, I, or II v. III), first-degree family history of breast cancer, annual household income ($; midpoints of household income categories), having at least a college degree (yes/no), having health insurance (yes/no), race (non-Hispanic white v. other), partner status at the time of diagnosis (being married/living with a partner v. divorced/separated/widowed/never married).

Statistical Analysis

We compared demographic and clinical characteristics by parental status (frequencies/chi-squared tests, and means/t tests) and proportions of women reporting different roles in treatment decisions using bar plots/chi-squared tests. We used univariate and multivariate logistic regression models to test associations between parental status and surgery type and receipt of chemotherapy, reporting odds ratios (ORs) and confidence intervals (CIs).

We ran several sensitivity analyses. First, we reviewed the medical records of Minnesota participants (accessible to the authors) who had consented to their medical record use, comparing their true parental status with our proxy measure. Second, we ran modified models: 1) excluding women aged ≤ 40 to whom chemotherapy may have been presented as clinically compulsory and 2) adjusting for potential confounders available in only a subset of participants (i.e., HER2 and nodal status; chemotherapy model) and having had a high-risk finding after genetic testing (surgery model).

Results

Of the 225 women aged 20 to 50 years, we classified 159 (70.7%) as having dependent children. Parents were more likely to be partnered and had higher incomes (P = 0.002 and 0.0003, respectively; Table 1). In the univariate logistic regression (data not shown), parents were more likely to have had bilateral mastectomy than unilateral mastectomy or breast-conserving surgery (with v. without dependent children, OR: 2.50, 95% CI: 1.38–4.54, P = 0.003). This association was stronger in the multivariate analysis (OR: 3.09, 95% CI: 1.44–6.62, P = 0.004). There was no evidence for differences in receipt of chemotherapy by parental status (with v. without dependent children, univariate OR: 0.69, 95% CI: 0.37–1.27, P = 0.23; multivariate OR: 0.75, 95% CI: 0.31–1.83, P = 0.53). Results from our sensitivity analyses supported these findings.

Table 1.

Baseline Characteristics of Participants Aged 20 to 50 Years, by Parental Status (N = 225)

	Dependent Children, n = 159	No Dependent Children, n = 66
Variable	Mean (SD)	Mean (SD)	P Value
Age at diagnosis, y	43.1 (5.0)	41.9 (6.8)	0.16
Annual household income, $1k USD	$92.7 ($36.9)a	$72.3 ($36.9)b	0.0003
		No. (%)
Race			0.09
White, Non-Hispanic	143 (89.9)	54 (81.8)
Other	16 (10.1)	12(18.2)
Education			0.2
No college degree	51 (32.1)	27 (40.9)
At least college degree	108 (67.9)	39 (59.1)
Marital status			0.002
Widowed/divorced/never married	20 (12.7)	20 (30.3)
Married/partnered	138 (87.3)	46 (69.7)
Health insurance			0.6
No	3 (1.9)	2 (3.0)
Yes	156 (98.1)	64 (97.0)
Breast cancer family history			0.39
No	113 (76.9)	51 (82.3)
Yes	34 (23.1)	11(17.7)
Cancer stage			0.97
0/I/II	104 (65.4)	43 (65.2)
III	55 (34.6)	23 (34.9)
Surgery			0.009
Breast conserving	51 (32.1)	31 (47.0)
Unilateral mastectomy	17 (10.7)	12 (18.2)
Bilateral mastectomy	91 (57.2)	23 (34.9)
Chemotherapy			0.23
No	67 (42.1)	21 (33.3)
Yes	92 (57.9)	42 (66.7)

^an = 148.

^bn = 60. The numbers highlighted in bold font in Table 1 indicate P-values < 0.05.

Women reported taking a more active decision role in surgery (Figure 1; P < 0.0001), with 50% reporting they made the decision with input from their doctors (v. 27% in chemotherapy), while 12% reported their doctors made the decision with input from them (v. 34% in chemotherapy). There were no substantial differences in these patterns by parental status (data not shown).

Figure 1.

Patient and physician roles in surgery and chemotherapy decision making among women aged 20 to 50 years (N = 225), Share Thoughts on Breast Cancer Study 2015.

A comparison of the parental status proxy measure with true parental status at diagnosis as verified from medical records for Minnesota participants revealed that we had excluded 2 of 25 true parents because they were aged >50 (i.e., outside our inclusion criteria). Among included Minnesota women, all parents aged 20 to 50 years were correctly identified (sensitivity: 100%, 95% CI: 85.2–100%). Two of 14 Minnesota women without dependent children were misclassified as parents (specificity: 85.7%, 95% CI: 57.2%–98.2%).

Discussion

We found that women with dependent children underwent bilateral mastectomy more often than similarly aged women without dependent children, but we found no similar differences in receipt of chemotherapy by parental status. Women reported having more active roles in surgery than in chemotherapy decision making, regardless of parental status.

More common receipt of bilateral mastectomy among parents in our study is consistent with previous reports of parents with cancer who stated that they would take more drastic treatment measures to prolong survival.^3,7-9 The lack of evidence for a similar association for chemotherapy suggests that chemotherapy decisions may differ. Some parents report that debilitating treatment side effects would be a reason for them to forego treatments.³ Another possible explanation, as supported by our data, is that surgery decisions may be more patient driven whereas chemotherapy decisions may be more provider driven. Given their young age and risk of recurrence, study participants may have seen little leeway to opt against chemotherapy which, unlike more flexible surgery options, may be felt to be a compulsory part of care. Prophylactic mastectomies were not our primary topic, but because bilateral breast cancer is rare,^11,12 many bilateral mastectomies in our study were likely prophylactic. Fears of recurrence and a sense of vulnerability and loss of control are consistently reported among women undergoing prophylactic mastectomy.^13-17 In addition, many women see surgery choices as their own, especially women with dependent children,¹⁸ and considering prophylactic mastectomy is associated with preferences to make one’s own treatment decisions.¹⁹ Women’s increasing sense of ownership in surgical decision making and a desire to regain a sense of control could be factors in recent trends toward more frequent prophylactic mastectomies.^20-22

A limitation of our study is the use of a proxy parental status measure, although our sensitivity analysis suggested that our measure had good sensitivity and specificity. Some participants may have had false recollections from the time of their diagnosis, but misclassification was likely rare because participants completed the study survey within 2 years of diagnosis. We had no information on the children’s age or individual physician recommendations, which may influence treatment decisions. HER2 and nodal status were not available for all participants, but a reduced model including HER2 and nodal status had similar findings as our main model. A limited sample size prevented us from testing for interactions of parental status with age. Observed treatment differences may reflect multiple factors, including preferences, socioeconomic status, costs, and unmeasured confounding. A strength of our analysis was our use of treatments received as outcomes, in contrast to hypothetical treatment scenarios.

Conclusions

To respect patients’ goals during their cancer experience, it is critical to acknowledge patients’ considerations about dependent children. Further research should assess differences in patients’ sense of ownership in treatment decision making by treatment type. Taking dependent children’s needs into consideration may be easier in surgery than, for example, in chemotherapy decisions.