Predictive Models of Psychological Distress, Quality of Life, and Adherence to Medication in Breast Cancer Patients: A Scoping Review

Figures & data

Figure 1 PRISMA Flow Diagram for the scoping review process.

Table 1 Original Research Studies on the Prediction of Psychological Distress: Characteristics and Results

Author (Year), Country	Cancer Type & Sample	Study Design	Predictive Model	Outcomes	Predictors	Results
Aquil (2021), Morocco^Citation32	Mixed female cancer patients (BC N 100 Age 48.26 SD= 11.40; Gynecological cancer N 100 Age 50.94 SD= 11.33)	Cross- sectional study	Multiple regression	Mental health disorders (anxiety, depression, body image dissatisfaction)	Socio-demographic and medical characteristics	Being younger (p<0.001), lower income (p<0.01), and advanced stage of the disease predicted mental disorders (p<0.01); (higher anxiety R^2 0.145 F=12.26; depression R^2 0.14 F=11.89; and greater body image dissatisfaction R^2 0.33 F=25.84).
Ganz (2003), USA^Citation38	N 635 Breast cancer female patients (Age>65)	Prospective cohort study	Logistic regression	Self-perceived health and psychosocial adjustment	Demographic, medical characteristics, QoL, and psychosocial variables (physical functioning, mental health, social support, medical interaction)	Predictors of better psychosocial adjustment (R^2 0.29; p=0.0001) were better mental health (p<0.001), emotional social support (p<0.001), and better self-rated interaction with healthcare providers (p<0.001).
Gõmez-Campelo (2014), Spain^Citation33	Mixed female cancer patients (BC N 50 Age 48.36 SD= 13.27; Gynecological cancer N 50 Age 48.82 SD= 14.43)	Cross- sectional study	Multiple regression	Psychological distress (body image disturbance, self-esteem, depression, and anxiety)	Socio-demographic characteristics, medical history, and social support	Being younger (p<0.001), inactive occupational status (p<0.05), and post-adjuvant therapy side effects (p<0.05) predicted body image disturbance (R^2 51.5% F=18.34) and depression (R^2 45.0% F=14.35). Anxiety (R^2 19.4% F=4.93) was predicted by post-adjuvant therapy side effects (p<0.01).
Haun (2021), Germany^Citation39	N 496 Mixed cancer patients (F= 269; Age 64.9 SD= 10.95)	Prospective, cross- sectional study	Machine learning approach	Anxiety	Socio-demographic and medical characteristics	Physical symptoms (fatigue/weakness) (β=0.181), Insomnia (β=0.122), and pain (β=0.041) predicted anxiety (LASSO regression: RMSE=0.370 R^2 0.415).
Hong (2020), Korea^Citation40	N 203 Mixed cancer patients (F=136; Age 57.3 SD=11.04)	Cross- sectional study	Structural equation modeling	Psychological adjustment	Symptom distress, contextual factors, Fear of Cancer Recurrence (FCR), and illness representations	Greater symptom distress, lower perceived social support and optimism, greater FCR, and negative perceptions of disease outcomes (ß=0.18, p<0.006; ß=0.35, p<0.001; ß=0.25, p<0.001 respectively) predicted poorer psychological adjustment (χ2=41.2, GFI=0.95, p<0.001).
Morasso (2001), Italy^Citation41	N 184 Breast cancer female patients (Age<70)	Randomized controlled study	Logistic regression, multivariate analysis	Psychiatric disorders	Clinical, demographic, and psychological variables (psychological distress, anxiety, depression)	From the logistic regression analysis, increasing age (p=0.001) and psychiatric history (p<0.001) were associated with psychiatric disorders at follow-up; The multivariate models showed that psychological distress predicted psychiatric disorders, while the predictive value of age and psychiatric history decreased from baseline to follow-up (AUC=77.0–91.8).
Schnoll (2002), USA^Citation30	N 109 Mixed cancer patients (F= 83; Age 60.3 SD= 11)	Cross-sectional study	Structural equation modeling	Adjustment (sexual, extended family, social, psychological)	Demographic, clinical, and psychosocial variables (meaning of life, optimism, coping strategies, social support)	Higher levels of social support and meaning in life (ß=0.38, p<0.05), and lower levels of avoidant- type coping behaviors (ß=0.77, p<0.05) predicted higher adjustment (χ2[81]=136.42, p<0.001; CFI=0.94; all factor loadings p’s<0.01).
Seow (2021), Canada^Citation31	N 255.494 Mixed cancer patients (F=145.632; Mean Age 64 SD= n.a.)	Retrospective, population-based, predictive study	Logistic regression	Poor performance status and severe symptoms (severe pain, dyspnea, wellbeing, and depression)	Demographic characteristics, clinical data, patient-reported outcomes, healthcare use	Lung disease, dementia, diabetes, radiation treatment, hospital admission, pain, depression, transitional performance status, issues with appetite, and receipt of home-care, increased by 10% the risk of low performance status and severe symptoms (including depression) (AUC=0.709–0.807).
Syrowatka (2018) Canada^Citation13	N 16.495 Breast cancer female patients (Age 65.7 SD= 11.8)	Population-based cohort study	Time-varying Cox proportional hazards models	New-onset psychological distress	Socio-demographic, breast cancer characteristics, and treatment received	Two predictive models were developed: (1) for the period of hospital-based treatment, and (2) for survivorship. Predictors of psychological distress in both periods were: younger age HR (95% CI) 0.91 (0.90–0.93) (1) 0.95 (0.94–0.97) (2), receipt of axillary lymph node dissection Impact at 6 months HR 0.98 (0.81–1.18) (1) 1.16 (1.05–1.28) (2), rheumatologic disease (at baseline HR 1.17 (1.07–1.28) (1) 1.23 (1.13–1.35) (2) or follow-up HR 1.34 (1.07–1.69) (1) 1.35 (1.12–1.63 (2), baseline menopausal symptoms impact at 6 months HR Proportional (1) 1.17 (1.05–1.32) (2), new opioid dispensations impact at 6 months HR 1.56 (1.29–1.87) (1) Proportional (2), emergency department visits HR 1.33 (1.21–1.47) (1) 1.47 (1.31–1.65) (2), and hospital contacts during follow-up HR 1.14 (1.03–1.25) (1) 2.32 (1.77–3.03) (2). In period (1) other predictors were: more advanced BC and chemotherapy Impact at 6 months HR 0.98 (0.81–1.18). In period (2) other predictors were: diagnosis of localized breast disease HR 1.15 (1.01–1.31), shorter duration of hospital-based treatment HR 0.94 (0.93–0.96), receipt of additional hospital-based treatment in survivorship HR 1.85 (1.33–2.59), and newly diagnosed comorbidities or symptoms (n.a.)*

Notes: *The only relevant metrics reported in the paper are HRs; since this is a measure of statistical association rather than a measure of predictive validity, the reported results should be considered with caution (see Limitations section).

Abbreviations: QoL, Quality of Life; SD, standard deviation; F, female; n.a., not available; CFI, comparative fit index; HR, hazard ratio; RMSE, root mean squared error; GFI, goodness of fit index.

Table 2 Original Research Studies on the Prediction of Quality of Life: Characteristics and Results

Author (Year), Country	Cancer Type & Sample	Study Design	Predictive Model	Outcomes	Predictors	Results
Liang (2022), China^Citation42	N 506 Breast cancer female patients (Age >18)	Prospective cohort study	Meta-based univariate regressions	QoL and survivorship	Resilience, age, income, tumor characteristics, and baseline QoL	Younger age, tumor characteristics, higher resilience, and baseline QoL predicted higher QoL outcomes after one year (P<0.0001 for all; C-indexes of 0.862 (95% CI, 0.815–0.909), 0.828 (95% CI, 0.745–0.910), 0.880 (95% CI, 0.816–0.944), and 0.869 (95% CI, 0.796–0.941) respectively).
Achimaș-Cadariu (2015), Romania^Citation27	N 51 Breast cancer female patients (Age 38.18 SD= 4.07)	Prospective randomized controlled study	Multiple linear regression	QoL	Demographic and psychological factors (emotional distress, perception of threat, anxiety, self-blame, positive reappraisal, and catastrophizing)	Emotional distress (β=−0.42, p<0.001) and catastrophizing as a coping strategy (β=−0.27, p=0.002) predicted poorer QoL. The model showed reasonable goodness of fit (F-test, p<0.001).
Ashing-Giwa (2007), USA^Citation43	N 703 Breast cancer female patients (Age 55 SD= n.a.)	Cross -sectional study	Multivariate regression analysis	QoL	Cancer-related medical factors, general health status, psychological (emotional wellbeing), demographic, healthcare system (quality of doctor-patient relationship), socio-ecological (social support) and cultural variables	Years since diagnosis, number of comorbidities, role limitations, emotional wellbeing, quality of doctor–patient relationship, social support, and life stress predicted QoL explaining 70% of the variability (R^2 0.08 F (6630) =29.0; p<0.001).
Carmona-Bayonas (2021), Spain^Citation44	N 339 Breast cancer female patients (F= 333; Age 52 SD= n.a.)	Prospective cohort study	Constrained partial proportional odds Bayesian predictive model	QoL	Age, medical variables, perceived risk of recurrence, and baseline QoL	Six covariates (chemotherapy regimen, TNM stage, axillary lymph node dissection, perceived risk of recurrence, age, type of surgery, and baseline EORTC scores) predicted QoL (C-index=0.65 (95% HPDI, 0.63–0.67); Brier score=0.21 (95% HPDI, 0.19–0.28).
Di Meglio (2022), France^Citation28	N 5000 Breast cancer female patients (Age 56.3 SD= 11.2)	Prospective longitudinal cohort multicenter study	Multivariate logistic regression model	Global fatigue; Secondary outcome: physical, emotional, and cognitive fatigue	Demographic, clinical and treated-related factors, and symptoms (anxiety, depression, insomnia, pain, hot flushes)	Younger age, higher BMI, smoking behavior, physical and psychological symptoms (fatigue, anxiety, insomnia, and pain) predicted severe fatigue two years after diagnosis (AUC=0.73 (95% CI, 0.72 to 0.75)).
Giedzinska (2004), USA^Citation29	N 621 Breast cancer female patients (Age 55.23 SD= 11.56)	Cross-sectional study	Linear regression analysis	QoL	Medical and demographic characteristics, cancer-related and psychosocial measures (depression, mental health, support, body image, sexual interest, vulnerability, emotional wellbeing)	Age, income, having had mastectomy, having had chemotherapy, and geographic location of participants predicted different aspects of QoL for different ethnic groups.*

Notes: *The paper report several different predictive values for each variable and for each ethnic group considered (ie, African Americans, Whites, Latinas, Asian Americans).

Abbreviations: QoL, Quality of Life; SD, standard deviation; F, female; n.a., not available; AUC, area under the receiver operating characteristic curve; HPDI, highest posterior density interval.

Table 3 Original Research Studies on the Prediction of Adherence and Treatment Discontinuation: Characteristics and Results

Author (Year), Country	Cancer Type & Sample	Study Design	Predictive Model	Outcomes	Predictors	Results
Cahir (2017), Ireland^Citation45	N 3415 Breast cancer female patients (Age 61.4 SD= 12.6)	Population- based cohort study	Multivariate relative risks model	Non- persistence	Demographic, clinical, and treatment-related risk factors	Women aged<50 years and prescribed antidepressants resulted at greater risk, while married women and those with previous use of medications had a lower risk. However, the model has limited predictive ability (AUC=0.61).
Henry (2012), USA^Citation46	N 503 Breast cancer female patients (Age 59 SD= n.a.)	Prospective randomized controlled study	Univariate and multivariate analysis of predictors	Treatment discontinuation	Demographics, clinical, and treatment-related characteristics	Younger age (HR, 1.4; 95% CI, 1.0 to 1.9; p=0.04), taxane-based chemotherapy (HR, 1.9; 95% CI, 0.99 to 3.6; p=0.048), and pre-existing pain (HR, 1.1; 95% CI, 1.0 to 1.2; p=0.04) predicted aromatase inhibitors discontinuation.*
Kuo (2022), Taiwan^Citation47	N 385 Breast cancer female patients (Age 55.1 SD= n.a.)	Longitudinal observational retrospective cohort study	Multiple logistic regression	Medication-taking behavior	Demographic, clinical, and treatment-related characteristics	The top five predictors as resulting from a Borda count based on three different models (MLR, AUC=1.00; DT, AUC=0.93; ANN, AUC=0.90) were: duration of AET discontinuation, duration of AET use, age at diagnosis (younger and older patients had lower adherence), BMI (lower BMI predicted poorer adherence), and radiotherapy; vitality status, regional lymph nodes positive, comorbidities, cause of death, and type of first recurrence were other predictors.
Meneveau (2020), USA^Citation48	N 11037 Breast cancer female patients (Age >70)	Retrospective cohort study	Stepwise selection and logistic regression	Adherence and medication initiation	Comorbidities, socio-economic measures, prescription medications, and demographics	Models were poorly predictive (AUC=0.65–0.60), demonstrating that the reasons for initiation and adherence to AET are complex and individual to the patient, and difficult to predict.
Shinn (2022), USA^Citation49	N 82 Breast cancer female patients (Age 53–54 SD= n.a.)	Prospective study	Discrete decisional logic applied to a behavioral feedback network	Treatment discontinuation	Patient-related factors, patient-provider relationship, treatment, and comorbid factors	Low risk perception, below-median QoL, low AET-related side effects, low level of cancer recurrence worry, and median levels of general anxiety characterized the profiles leading to AET discontinuation in the simulations conducted with the models.**
Yanez (2021), USA^Citation50	N 954 Breast cancer female patients (Age 56.6 SD= 8.9)	Randomized clinical study (post hoc analysis)	Cox proportional hazards regression	Treatment discontinuation	Medical comorbidities and QoL variables	Depression (HR, 1.82; 95% CI, 1.19–2.77; p=0.005) and poor social (HR, 1.94; 95% CI, 1.20–3.13; p=0.006) and physical wellbeing (HR, 2.12; 95% CI, 1.30–3.45; p=0.002) are significant risk factors for early discontinuation of AET.*

Notes: *The only relevant metrics reported in the paper are HRs; since this is a measure of statistical association rather than a measure of predictive validity, the reported results should be considered with caution (see Limitations section). **Results based on repeated logic network simulations conducted with 96 candidate decisional models over a broad range of initial demographic and behavioral profiles. No predictive values are available.

Abbreviations: QoL, Quality of Life; SD, standard deviation; n.a., not available; AET, adjuvant endocrine therapy; AUC, area under the receiver operating characteristic curve; HR, hazard ratio; CI, confidence interval; MLR, multiple logistic regression; DT, decision tree; ANN, artificial neural network.

Aquil A, El Kherchi O, El Azmaoui N, et al. Predictors of mental health disorders in women with breast and gynecological cancer after radical surgery: a cross-sectional study. Ann Med Surg. 2021;65:102278. doi:10.1016/j.amsu.2021.102278

 PubMed Web of Science ®Google Scholar

Ganz PA, Guadagnoli E, Landrum MB, Lash TL, Rakowski W, Silliman RA. Breast cancer in older women: quality of life and psychosocial adjustment in the 15 months after diagnosis. J Clin Oncol. 2003;21(21):4027–4033. doi:10.1200/JCO.2003.08.097

 PubMed Web of Science ®Google Scholar

Gõmez-Campelo P, Bragado-álvarez C, Hernández-Lloreda MJ. Psychological distress in women with breast and gynecological cancer treated with radical surgery. Psychooncology. 2014;23(4):459–466. doi:10.1002/pon.3439

 PubMed Web of Science ®Google Scholar

Haun MW, Simon L, Sklenarova H, Zimmermann-Schlegel V, Friederich H-C, Hartmann M. Predicting anxiety in cancer survivors presenting to primary care - A machine learning approach accounting for physical comorbidity. Cancer Med. 2021;10(14):5001–5016. doi:10.1002/cam4.4048

 PubMed Web of Science ®Google Scholar

Hong SJ, Shin NM, Jung S. A predictive model of fear of cancer recurrence for patients undergoing chemotherapy. Support Care Cancer. 2020;28:4173–4181. doi:10.1007/s00520-019-05245-7

 PubMed Web of Science ®Google Scholar

Morasso G, Costantini M, Viterbori P, et al. Predicting mood disorders in breast cancer patients. Eur J Cancer. 2001;37(2):216–223. doi:10.1016/S0959-8049(00)00390-7

 PubMed Web of Science ®Google Scholar

Schnoll RA, Knowles JC, Harlow L. Correlates of adjustment among cancer survivors. J Psychosoc Oncol. 2002;20(1):37–59. doi:10.1300/J077v20n01_03

 Web of Science ®Google Scholar

Seow H, Tanuseputro P, Barbera L, et al. Development and validation of a prediction model of poor performance status and severe symptoms over time in cancer patients (PROVIEW+). Palliat Med. 2021;35(9):1713–1723. doi:10.1177/02692163211019302

 PubMed Web of Science ®Google Scholar

Syrowatka A, Hanley JA, Weir DL, Dixon WG, Meguerditchian AN, Tamblyn R. Ability to predict new-onset psychological distress using routinely collected health data: a population-based cohort study of women diagnosed with breast cancer. JNCCN J Natl Compr Cancer Netw. 2018;16(9):1065–1073. doi:10.6004/jnccn.2018.7038

 PubMed Web of Science ®Google Scholar

Liang MZ, Tan Y, Knobf MT, et al. Resilience index improves prediction of 1-year decreased quality of life in breast cancer. J Cancer Surviv. 2022;2022:1–10.

 Google Scholar

Achimaș-Cadariu P, Iancu M, Pop F, Vlad C, Irimie A. Psychological screening and health related quality of life in Romanian breast cancer survivors. J Evid Based Psychother. 2015;15(2):267.

 Web of Science ®Google Scholar

Ashing-Giwa KT, Tejero JS, Kim J, Padilla GV, Hellemann G. Examining predictive models of HRQOL in a population-based, multiethnic sample of women with breast carcinoma. Qual Life Res. 2007;16:413–428. doi:10.1007/s11136-006-9138-4

 PubMed Web of Science ®Google Scholar

Carmona-Bayonas A, Calderón C, Hernández R, et al. Prediction of quality of life in early breast cancer upon completion of adjuvant chemotherapy. NPJ Breast Cancer. 2021;7(1):92. doi:10.1038/s41523-021-00296-8

 PubMed Web of Science ®Google Scholar

Di Meglio A, Havas J, Soldato D, et al. Development and validation of a predictive model of severe fatigue after breast cancer diagnosis: toward a personalized framework in survivorship care. J Clin Oncol. 2022;40(10):1111–1123. doi:10.1200/JCO.21.01252

 PubMed Web of Science ®Google Scholar

Giedzinska AS, Meyerowitz BE, Ganz PA, Rowland JH. Health-related quality of life in a multiethnic sample of breast cancer survivors. Ann Behav Med. 2004;28(1):39–51. doi:10.1207/s15324796abm2801_6

 PubMed Web of Science ®Google Scholar

Cahir C, Barron TI, Sharp L, Bennett K. Can demographic, clinical and treatment-related factors available at hormonal therapy initiation predict non-persistence in women with stage I–III breast cancer? Cancer Causes Control. 2017;28:215–225. doi:10.1007/s10552-017-0851-9

 PubMed Web of Science ®Google Scholar

Henry NL, Azzouz F, Desta Z, et al. Predictors of aromatase inhibitor discontinuation as a result of treatment-emergent symptoms in early-stage breast cancer. J Clin Oncol. 2012;30(9):936. doi:10.1200/JCO.2011.38.0261

 PubMed Web of Science ®Google Scholar

Kuo CC, Wang HH, Tseng LP. Using data mining technology to predict medication‐taking behaviour in women with breast cancer: a retrospective study. Nursing Open. 2022;9(6):2646–2656. doi:10.1002/nop2.963

 PubMed Web of Science ®Google Scholar

Meneveau MO, Keim-Malpass J, Camacho TF, Anderson RT, Showalter SL. Predicting adjuvant endocrine therapy initiation and adherence among older women with early-stage breast cancer. Breast Cancer Res Treat. 2020;184:805–816. doi:10.1007/s10549-020-05908-8

 PubMed Web of Science ®Google Scholar

Shinn E, Busch BE, Jasemi N, et al. Network modeling of complex time-dependent changes in patient adherence to adjuvant endocrine treatment in ER+ breast cancer. Frontiers in Psychology. 2022;2022:4172.

 Google Scholar

Yanez B, Gray RJ, Sparano JA, et al. Association of modifiable risk factors with early discontinuation of adjuvant endocrine therapy: a post hoc analysis of a randomized clinical trial. JAMA Oncol. 2021;7(8):1196–1202. doi:10.1001/jamaoncol.2021.1693

 Web of Science ®Google Scholar