Background

Navigating a cancer diagnosis can be distressing. For many patients with gynecologic cancer, anxiety and distress begin prior to a formal cancer diagnosis, with the nebulous worries of symptoms such as vaginal bleeding or vague abdominal pain.¹

Since 1997, NCCN has recognized the importance of screening for and managing distress.² Routine screening for distress, preferably at periods of increased vulnerability, is recommended by NCCN and endorsed by the National Academy of Medicine, the American College of Surgeons, and the Oncology Nursing Society.^2,3 Distress has been described as the sixth vital sign and is often underrecognized by clinicians.^2,4–6 Increased levels of distress have wide-ranging effects, including decreased adherence to cancer treatment, longer hospital stays, postoperative complications, worsened quality of life, and decreased overall survival.^6–12

For some patients, cancer-related distress may be transient and manageable without formal intervention, whereas for others, it can persist throughout treatment and survivorship.^6,13

When compared with other cancer groups, patients with gynecologic cancer experience some of the highest rates of distress, with up to 85% reporting elevated levels of anxiety and up to 54% reporting symptoms of depression.^{5,7,10,12–14} Additionally, female patients with cancer have been shown to experience anxiety and depression at twice the rate of their male counterparts.⁵

Distress has been assessed using a wide range of tools.^{2,5,8,12,15,16} Endorsed by NCCN, the Distress Thermometer (DT) has been validated in patients with cancer in general as well as in patients with gynecologic cancer specifically.^2,6,16 Despite a burgeoning body of evidence surrounding distress in gynecologic cancer, gaps in knowledge exist regarding both risk factors for increased distress and optimal timing of collection. Many existing studies focus on survivors of gynecologic cancers as opposed to patients actively receiving treatment.^15,17–19 Other studies limit participation to patients with early-stage disease.^15,20

At our institution, the University of Wisconsin-Madison, DTs are administered at all postoperative visits, which has been identified as a vulnerable time for patients with cancer as they await their final pathology results.^2,21 The objective of this study was to explore risk factors associated with distress, as measured by the NCCN DT, among a large cohort of patients presenting to a postoperative visit at a gynecologic cancer center.

Patients and Methods

A retrospective study was performed to examine levels of distress among postoperative patients at a single academic gynecologic oncology practice. Distress levels were examined in conjunction with disease-specific, medical, and sociodemographic variables to identify associations. Distress was measured using NCCN DT numerical results, obtained from all postoperative patients in the practice from September 2016 through December 2020. In our practice, patients are seen 2 weeks after surgery, at which time final pathology results are reviewed. DTs were administered at the beginning of the postoperative visit, prior to patients being seen by the attending physician. Patients were excluded from analysis if they did not have a documented NCCN DT result from the time of their postoperative visit or if planned variables for analysis were missing from the electronic medical record (EMR). This study was considered exempt under the Institutional Review Board under a quality improvement exemption.

Patient data were collected retrospectively on women who completed a postoperative visit in our gynecologic oncology clinics. Patients with missing variables after utilization of the EMR dashboard had a detailed chart review performed by the research team to identify missing variables.

Variables included in the analysis were age, body mass index (BMI), payer status, partner status, smoking status, alcohol use, substance use, disease site, procedure type, and medical comorbidities representative of common comorbidities in our patient population, including asthma, depression, diabetes, and hypertension.

Disease site was assigned based on the presumed diagnosis at the time of discharge from surgical admission, incorporating preoperative diagnosis, intraoperative findings, and frozen section. We chose to categorize disease site in this way to reflect the working diagnosis that the patient would have at the time of DT recording. Benign designation was given to all patients without a cancer diagnosis prior to surgery in whom working diagnosis at the time of discharge from the hospital was a likely benign process. Patients in whom malignancy could not be reasonably ruled out at the time of discharge were coded by disease site and probable malignancy. The prophylactic category reflected all patients undergoing risk-reducing or prophylactic surgery without preoperative concern for malignancy. The “other” category consisted of diagnoses not fitting described categories and included gestational trophoblastic disease, vaginal, bowel, obstetric (abdominal cerclage), Ewing sarcoma of the bone, pelvic abscess, and pelvic schwannoma.

Primary outcome was the NCCN DT numerical result. The DT contains a graphical depiction of a thermometer accompanied by a definition of distress on a scale of 0 (no distress) to 10 (extreme distress) (Figure 1). At our institution, DTs are administered to patients by medical assistants and are filled out at the beginning of every 2-week postoperative visit. Numerical results are entered into the EMR, and results are reviewed by clinicians. As recommended by NCCN, patients with a score <4 (low distress) have specific concerns addressed by the clinician and are not routinely referred for supportive services.² Patients with clinically significant distress (scores ≥4) are routinely referred for supportive services. In our practice, patients with a score of 4 to 8 (moderate distress) are referred to a clinic social worker, who performs a needs assessment and provides a resource recommendation, which may include referral to a sexual health clinic, palliative care services, mental health services, or financial assistance services. Patients with a score of 9 to 10 (high distress) are referred urgently and receive same-day contact by social work services in order to assess the need for immediate supportive services or referrals.

View Full Size

NCCN Distress Thermometer.

Reproduced with permission from the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines^®) for Distress Management V.1.2024. © 2023 National Comprehensive Cancer Network, Inc. All rights reserved. Accessed March 13, 2023. To view the most recent version, visit NCCN.org.

Citation: Journal of the National Comprehensive Cancer Network 22, 2; 10.6004/jnccn.2023.7093

NCCN DT results were analyzed as both continuous variables from 0 to 10 and as discretized values. Discretized values (DT ≤3 = low distress; DT 4–8 = moderate distress; or DT ≥9 = high distress) were chosen based on the clinic thresholds for distress score intervention previously described. Age and BMI were reported as a mean with standard deviation. Univariate and multivariate regression analyses were performed to analyze differences in NCCN DT results by multiple personal, demographic, and disease-related variables. Risk factors were determined using the Wald test with Bonferroni correction for univariate analyses. First, we fit linear regression models by treating the original distress score as a continuous, numeric variable. Second, we fit logistic regression models by discretizing the distress score into 2 categories: low distress (DT ≤3) versus moderate/high distress (DT ≥4). We chose to combine all scores ≥4 into one category of moderate/high distress because this is consistent with the recommendations for referral to supportive services in the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for Distress Management² and is also consistent with referral cutoffs reported in previous studies.^7,15,21 The threshold for statistical significance was P<.05. On univariate analysis incorporating Bonferroni correction, the P value for significance at a level of 0.1 was .002.

Results

After 79 initial charts were excluded due to missing variables, a total of 1,874 patients were identified for analysis. Nine patients did not have a completed DT and 70 did not have a documented payer status and were therefore excluded. The final analysis included NCCN DT data from 1,795 postoperative visits (Figure 1).

Mean [SD] age was 58.30 [13.99] years and mean [SD] BMI was 31.50 [9.13] kg/m². The majority (n=1,663; 92.64%) of patients identified as non-Hispanic White. Given the homogeneity of population, race and ethnicity were not included for statistical analysis (Table 1).

Table 1.

Demographic and Clinical Variables Associated With Distress (N=1,795)

View Table

Discharge diagnosis distribution demonstrated that most patients had uterine (n=677; 37.72%) or ovarian (n=650; 36.21%) primary disease sites. Disease sites making up <1% of the total were grouped as “other” and included primary vaginal disease, gestational trophoblastic disease, and bowel surgery. Of all patients presenting to a postoperative visit, 236 (13.15%) had knowledge of presumed benign pathology prior to completion of the NCCN DT (Table 1).

Type of procedure was categorized as laparoscopy (n=956; 53.26%), laparotomy (n=609; 33.93%), radical vulvar (n=72; 4.01%), and minor (n=158; 8.8%). The “minor” category included hysteroscopy/dilation and curettage, colposcopy, laser therapy, cervical excision, cystoscopy, examination under anesthesia, biopsies, and wide local excision.

Medical comorbidities were common, with 28.69% (n=515) of patients reporting a diagnosis of hypertension, 11.64% (n=209) reporting a history of diabetes, 10.31% (n=185) with diagnosis of depression, and 7.08% (n=127) with a history of asthma. More than half of patients (n=1,060; 59.05%) reported a partnered relationship status and had private insurance (n=1,019; 56.77%) (Table 1).

Most patients (n=1,288; 71.75%) endorsed low levels of distress, 26.41% (n=474) reported moderate levels of distress, and a minority of patients (n=33; 1.84%) documented a high distress level. Some distress was documented in most patients, with 70.42% (n=1,264) reporting a score >0 (Table 2).

Table 2.

Distress Score Distribution With Discretized Values (N=1,795)

View Table

On univariate analysis incorporating Bonferroni correction (Table 1), statistically significant associations were seen between distress as a continuous variable and age (P=.002), and depression (P<.001), and current smoking status (P=.004). When assessing distress as discretized values based on clinical thresholds for referrals, univariate relationships were significant for depression (P≤.001) and status as a current smoker (P=.001).

On multivariate analysis (Table 1), with distress as a continuous variable, significant associations were seen between age (P=.006), asthma (P=.041), depression (P<.001), status as a current smoker (P=.028), and knowledge of benign disease (P=.002). On multivariate analysis, with distress as discretized values, significant variables were depression (P<.001) and status as a current smoker (P=.005).

Disease site and procedure type were not significantly associated with levels of distress in any model (Table 1).

Discussion

In this retrospective review of nearly 1,800 patients in the postoperative period, distress levels were overall low, although most patients did report some level of distress. Slightly more than one-fourth of all patients had clinically significant distress, with a DT score ≥4. This is consistent with results from comparative studies.^{5,7,8,16,22,23}

Our sample included all patients presenting postoperatively after undergoing surgery with our gynecologic oncology providers and included a subset (13%) of patients with presumed benign disease at the time of NCCN DT completion. Although the postoperative period has been identified as a period of elevated distress, previous studies have excluded patients with benign pathology.²² We chose to include patients with benign pathology because this group makes up a significant proportion of patients receiving surgery and postoperative care in gynecologic oncology centers.^23,24 Often, this is due to concern for possible malignancy prior to surgery. This uncertainty, as well as the known multifactorial nature of distress, could potentially be a source of anxiety and distress.^2,23 Given the lack of prior evaluation of effect of benign status on distress, we believed it to be important to include this subgroup in the analysis. In our study, patients with knowledge of probable benign disease at the time of discharge from surgery had significantly decreased distress.

With procedure type stratified into 4 categories—laparoscopy, laparotomy, minor, and radical vulvar—no association was seen with distress levels. This relationship has not been previously described, although prior studies have consistently documented increased levels of distress among patients experiencing treatment-related sequelae, such as lymphedema, surgical menopause, or infertility.^14,18,20,25 We similarly saw no correlation between disease site and distress results, which corroborates findings from most existing studies.^{12,15,22,26–28} These results, along with our findings of decreased distress among patients with benign disease, indicate that increased postoperative distress may stem from factors directly related to processing and navigating a new cancer diagnosis rather than surgical intervention alone.²¹

Our findings highlight the intertwining effects of gynecologic oncology surgery experienced by some patients, such as loss of reproductive organs, possible surgical menopause, and the effect on future fertility, among others. Both young age and a history of depression were significantly associated with distress in our findings. This echoes findings of previous studies, which describe increased distress in patients with a history of depression.^12,26 Female patients with cancer and patients with gynecologic malignancies have shown an increased prevalence of clinically significant anxiety and depression, confirming that they are a particularly vulnerable population.^{5,6,10,12–14,28–31} Personal history of depression showed the strongest association with increased distress in our population and was significantly associated with distress in all models. Screening for depression early in the clinical relationship is crucial for identification of and early intervention in patients at risk for experiencing clinically significant distress.

Age also demonstrated a significant inverse relationship with levels of reported distress in both univariate and multivariate analyses when analyzed as a continuous variable. This finding correlates with most previously described relationships between age and distress in the literature.^{5,12,13,20,32} The known association between the severity of menopausal symptoms, infertility concerns, and levels of distress may explain some of this association, because younger women are more likely to experience significant changes in these domains during or after treatment of gynecologic malignancies compared with their older counterparts.^14,18 Discussion of these anticipated consequences may be especially relevant in younger patients presenting for gynecologic cancer care, with consideration of early integration of reproductive counseling, sexual health, and well-being supportive services.

Sociodemographic variables in our study, including partner and insurance status, were not associated with distress. These findings indicate that fixed variables such as partner status and insurance status may not reliably reflect unmet needs such as social support and financial distress, because financial toxicity and unmet social needs have consistently correlated with increased levels of distress and decreased resilience.^7,30,33 Personalized inquiries into these domains, instead of reliance on insurance and partner status, may yield a more fruitful understanding of stressors.

Strengths of this study include a large sample size with a prolonged period of data collection spanning 5 years. Additionally, consistent timing of data collection is a relative strength; all DT results are collected at the postoperative visit, and patients are consistently screened at the same time during the visit. Other strengths include a large number of variables analyzed, allowing identification of multiple risk factors and robust multivariate analysis. Use of discretized levels of distress, which correlate with recommended clinical thresholds for referral, increases internal and external validity. Additionally, although some previous studies have used a score ≥4 as a single cutoff point for referral services, many patients referred for such services do not pursue them.^4,27 Inclusion of a “high” level of distress, as done in our practice, has been suggested as an alternative and may be clinically useful for triaging urgency of referral for supportive services; future studies could evaluate the impact of this differential triaging system.^4,27,34

Although the timing of data collection is a strength of this study in some ways, it may also be seen as a limitation. Routine collection for all patients at the postoperative visit decreases selection bias; however, patients with initially an advanced-stage diagnosis, who may not be candidates for surgery, were not included. These patients, with more limited treatment options and poorer prognosis, have reported higher supportive care needs in survivorship, but have not consistently reported higher distress in previous studies, when compared with patients with surgically amenable disease.^12,13,26 Additionally, a cross-sectional analysis of distress at the postoperative visit does not allow us to track variation in distress over time.^21,28 Although logistically practical, the administration of DTs prior to discussion of surgical pathology results, postoperative concerns, and plans moving forward may affect the generalizability and utility of DT results; previous studies have demonstrated that the mere act of being asked about their distress, filling out the thermometer, and receiving a referral for services can lower levels of distress.^4,27 Additionally, this study had a homogeneous population; lack of ethnic and racial diversity is a limitation of the study.

We chose not to include final stage as a variable for analysis due to the timing of NCCN DT administration in our practice. Patients fill out the thermometer prior to meeting with the physician to review pathology and discuss plans for treatment. As such, many patients do not know their stage at the time of thermometer completion. This is consistent with recommendations in the NCCN Guidelines and is consistent with descriptions of practice in other studies, in which the DTs are administered in the waiting room prior to a visit or in the visit room during the nurse assistant or nurse check-in.^2,4

Conclusions

More than one-fourth of women undergoing surgery in a gynecologic oncology practice reported moderate or high distress at their postoperative visit. Understanding risk factors provides opportunities to identify patients most at risk for experiencing high distress. Although patients with known benign pathology demonstrated decreased distress, disease site and procedure type were not associated with distress, indicating that surgical intervention alone does not significantly impact levels of distress.

The association between unmodifiable risk factors and distress, such as young age, provides an opportunity for targeted intervention early in the treatment course. The association with modifiable risk factors such as personal history of depression and status as a current smoker provides insight into counseling and prevention strategies.

Future studies could investigate the change in reported distress among postoperative patients when surveyed at the beginning of the visit versus after the discussion of pathology and plan with the clinician, as well as evaluate differential cutoffs for referral to services, in an effort to improve specificity of the DT to capture patients with clinically relevant levels of distress. Such studies may improve the understanding of the link between elevated reported levels of distress and acceptance of referral to supportive services. Further research should continue to explore the utility of the DT in diverse patient populations.