You are here

Immediate versus delayed self-reporting of symptoms and side effects during chemotherapy: Does timing matter?

European Journal of Oncology Nursing, 2, 15, pages 130 - 136



The aim of this study was to examine how patients recall symptoms at a delayed self-report. Accurate insight into toxicity symptoms during chemotherapy is essential so that nurses and doctors can assess therapeutic tolerance and adjust supportive care accordingly.


A non-experimental, longitudinal design was employed. Using the Therapy-Related Symptoms Checklist (TRSC), respondents (n = 142) reported their initial symptoms during the first 7 days of the chemotherapy cycle at two different times: (1) each day of the first seven days after the chemotherapy administration (immediate self-report), and (2) at their next hospital visit for chemotherapy (delayed self-report). We compared the number and severity of symptoms and side effects reported in the immediate and delayed self-reports.


Respondents reported significantly fewer symptoms and fewer severe symptoms in the delayed self-report. For 22 out of 25 symptoms the delayed-reported grade was significantly lower than the immediate-reported maximum grade. Compared to the immediate-reported median grade, significant differences occurred in only 10 out of the 25 symptoms. In all cases, except fatigue, the delayed-reported grade was significantly higher than the immediate-reported median grade.


This study indicates that delayed self-report of chemotherapy side effects is not an appropriate measure of actual symptoms and side effects experienced by patients. Delayed self-report gives a weaker insight into actual symptom burden. Fatigue is at particular risk to be minimized at the delayed self-report. Therefore it is recommended to assess chemotherapy-related symptoms and side effects by means of immediate self-report.

Keywords: Cancer patients, Chemotherapy, Side effects, Symptom assessment, Self-report.


Obtaining accurate insight into the side effects experienced by chemotherapy patients is of great importance. Information on toxicity-related symptom burden can alter a patient’s need for dose modifications, supportive care, and education (Basch et al, 2005 and Basch et al, 2007). Ultimately, these interventions not only will influence patients’ side effects and quality of life but also will reduce the burden of cancer therapy ( Butt et al., 2008 ).

Many authors have discussed the importance of patient self-reports in the assessment of adverse effects of chemotherapy. Some studies have emphasized the following uses for self-report methods: to foster patient–clinician communication, to provide additional information about often-missed symptoms, and to improve the well-being and satisfaction of patients (Basch et al, 2005 and Basch et al, 2007). Some studies have relied on patient self-reports, as other methods are lacking. Indeed, shortened hospital stays have strongly reduced opportunities for caregivers to supervise and assess the effects of cancer and its treatment, forcing patients to take on a more active role in monitoring and communicating their side effects ( Kearney et al., 2006 ). Other reasons for using patient self-reports include the following: caregivers lack the time needed to carefully assess symptom burden, patients are uncomfortable in discussing some side effects face-to-face, and patients are unable to recall symptoms when questioned by caregivers ( Youngblood et al., 1994 ).

Additional arguments advocating the use of self-report methods can be found in studies comparing patient versus healthcare professional assessment of side effects. These studies report that professionals underestimate both the number and severity of side effects, and thus have recommended the use of patient self-report methods to overcome the inaccurate assessments of clinicians and nurses in detecting and reporting adverse effects (Basch et al, 2006, Braud et al, 2003, Fromme et al, 2004, Stromgren et al, 2001, and Youngblood et al, 1994). Agreement between the assessments of professionals and patients is better for more observable and physical symptoms than for more subjective and psychosocial symptoms ( Basch et al., 2006 ). This discrepancy between patient and healthcare professional assessment can be explained, in part, by the use of open-ended method assessments—which fail to capture aspects of the patients’ reactions to the toxic effects of chemotherapy—rather than structured paper questionnaires (Basch et al, 2006, Homsi et al, 2006, and Youngblood et al, 1994). Homsi et al. (2006) found, for example, a tenfold difference between the number of symptoms obtained voluntarily through open-ended questions and those obtained through systematic assessment. Researchers hypothesized that patients’ reluctance in volunteering symptoms can be explained by the perception that these symptoms are inevitable, that supportive care is unavailable or ineffective, that good patients do not complain, or that talking about symptoms would distract physicians from curing the cancer (Homsi et al, 2006 and Youngblood et al, 1994). Concerning the qualitative discrepancy between physician and patient symptom assessments, Basch et al. (2006) found only small and rarely clinically meaningful discrepancies on the severity of symptoms, when using a structured instrument such as the US National Cancer Institute Common Terminology Criteria for Adverse Events (CTCAE). Taken together, these findings support the use of structured instruments for symptom assessment. It is also clear that the patient self-report is a reliable method for assessing symptoms related to chemotherapy treatment ( Basch et al., 2006 ).

Healthcare professionals are not the only ones to have been studied as potential assessors of symptoms. Several studies have examined self-reports completed by the next of kin. A review by Lobchuk and Degner (2002b) , for example, revealed that family caregivers tend to over-report symptoms. As with healthcare professionals, better levels of agreement occurred for more objective and observable symptom experiences and for physical symptoms compared to psychosocial symptoms. Moreover, a considerably greater level of agreement was found on a global than on an individual symptom level (Lobchuk and Degner, 2002a and Lobchuk and Degner, 2002b).

In contrast to the reliability of the source (e.g., patient versus healthcare professionals) and symptom assessment method (e.g., open-ended versus structured), thus far the impact ofwhenthe symptom assessment takes place has received little research. The recommendation of written and structured symptom self-report based on earlier cited research does not resolve the matter ofwhenthis self-report should be asked. With respect to assessing chemotherapy-related symptoms and side effects, accurately gauging symptom burden and evolution throughout the entire chemotherapy cycle is as important as obtaining information about symptoms experienced at the time of a patient’s hospital visit. Youngblood et al. (1994) found that patients have difficulties recalling symptoms when healthcare professionals question them. Current practice of delayed symptom report at the research centre strengthens the assumption that asking patients to recall symptoms experienced earlier during treatment leads to underreporting or minimizing of the experienced symptom burden. Thus, the aim of the present study was to examine the correlation and the difference between immediately reported symptoms (immediate self-report) and symptoms recalled at a later time during the next hospital visit (delayed self-report), when using a self-report questionnaire. The underlying assumption of this study is that, for good therapeutic policy and supportive care, caregivers need accurate insight into the worst symptom burden experienced by patients. In addition it is hypothesized that, during a delayed self-report, patients tend to report on their symptom burden by averaging out between good and bad days.


Design and sample

This descriptive study was conducted at the University Hospitals Leuven, Belgium, in 8 oncology wards of which 2 are one-day clinics. Patients were recruited on 15 business days during 3 consecutive weeks, between May 25, 2009 and June 12, 2009. During the 3 subsequent weeks, recruitment continued for those patients who were either missed by the researchers or have had delay of treatment in the first recruitment weeks. The following inclusion criteria were used:

  • a) Adult oncology patient (18 years and older),
  • b) Speak and understand Dutch,
  • c) Able to fill out the symptom checklist,
  • d) Admitted to the hospital for the administration of a one-day chemotherapy protocol:
    • a. hospital stay no longer then a day and 1 overnight stay
    • b. next admission for chemotherapy administration at day one of the next chemotherapy cycle
  • e) No contra-indication made by the nurse in charge of the patient to invite this patient to participate,
  • f) Agree to participate in the study by signing the informed consent.

The study was approved by the Ethics Committee of the Catholic University Hospitals of Leuven. Researchers were provided with a list of scheduled chemotherapy patients by the ward managers of the participating wards and tracked eligibility of these patients in the electronic patient file. Eligible patients were invited to participate by one of the researchers, were explained about the study and were handed the information form attached to the informed consent. The researchers returned to the patients at the same day to find out their decision. For 6 patients, invitation to participate was contra-indicated by the nurse in charge for following reasons: ‘patient being too emotional’ (n = 1), ‘patient being too emotional after bad news’ (n = 2), ‘patient in too bad physical condition’ (n = 1), ‘study participation too burdensome in social context’ (n = 1). We identified 176 patients those were eligible for inclusion in the study. Eleven patients refused to participate for several reasons: ‘study being too incriminating’ (n = 2); ‘not feeling sick from chemotherapy and therefore considering study participation as useless’ (n = 2); ‘not in the mood to fill out questionnaires’ (n = 4), ‘not wanting to be focused on symptom burden’ (n = 2) and ‘considering the whole study as ridiculous’ (n = 1). Three patients did not return their questionnaires from the first round. In addition 20 patients did not complete a self-report questionnaire in the second round (delayed self-report). In 2 of these cases, this drop-out was a result of ‘not wanting to continue participation’. Other reasons were ‘cancellation of hospital appointments’ (n = 5), ‘next appointment being missed by the researchers’ (n = 8), ‘late self-report questionnaire completed but not received by the researcher’ (n = 3) and ‘incorrect inclusion and therefore exclusion from the study’ (n = 2). The final study sample consisted of 142 patients (response rate of 80.7%).

The majority of included patients were between 50 and 59 years old (36.6%), female (67.6%) and married (75.4%) ( Table 1 ). Only 31.7% hadn’t had earlier cancer treatment, 41.6% had already been treated with chemotherapy in the past. 16.2% of the respondents had the first cycle of the chemotherapy protocol when included in the study. 21.8% had already had more than 6 cycles of the protocol. Breast cancer and tumors in the digestive tract occurred the most, respectively in 33.8 and 30.3% of the respondents. There was no significant difference in sex between the responders and the non-responders. However, non-responders were significantly older than the responders (p < 0.05).

Table 1 Sample characteristics.

Characteristic n (%)
 Female 96 (67.6)
 Male 46 (32.4)
 20–29 2 (1.4)
 30–39 11 (7.7)
 40–49 18 (12.7)
 50–59 52 (36.6)
 60–69 33 (23.2)
 70–79 23 (16.2)
 80–89 3 (2.1)
Marital status
 Married 107 (75.4)
 Cohabitating 10 (7.0)
 Single 19 (13.4)
 Widowed 6 (4.2)
Earlier cancer treatment
 None 45 (31.7)
 Surgery alone 28 (19.7)
 Radiotherapy alone 2 (1.4)
 Chemotherapy alone 14 (9.9)
 Combination containing chemotherapy 45 (31.7)
 Combination without chemotherapy 8 (5.6)
Cycle in the chemotherapy protocol
 First 23 (16.2)
 Second 30 (21.1)
 Third 18 (12.7)
 Fourth 18 (12.7)
 Fifth 14 (9.9)
 Sixth 8 (5.6)
 More than sixth 31 (21.8)
Tumor site
 Breast 48 (33.8)
 Respiratory tract 12 (8.5)
 Head and neck 2 (1.4)
 Gynecologic organs 15 (10.6)
 Digestive tract 43 (30.3)
 Urologic organs 7 (4.9)
 Hematologic 10 (7.0)
 Other 5 (3.5)

Instruments and procedure

Demographic and clinical variables (age, sex, marital status, tumor site, earlier treatment, chemotherapy protocol, chemotherapy cycle) were collected through a short questionnaire. When answers on clinical questions were unavailable or unclear, they were corrected upon the electronic patient file. A validated symptom checklist, the Therapy-Related Symptoms Checklist (TRSC), was used for collecting self-report data on toxicity symptoms (Williams et al, 2001 and Williams et al, 2006). The TRSC is a structured self-report instrument, listing 25 possible toxicity symptoms of chemotherapy. The severity of symptoms is graded on a 5-point Likert scale (0 = none; 4 = very severe). An acceptable internal consistency (Cronbach’sα = 0.70) as well as a good construct and discriminant validity for the TRSC has been demonstrated ( Williams et al., 2001 ). Translation and back-translation were used to provide patients with a Dutch checklist: the instrument was first translated from the source language into the target language, and then back-translated in the source language, by 2 different native speakers. Since the 2 source language versions were quasi-identical, the use of the Dutch translation was considered justified. Since in a pilot study with 32 patients, several patients reported ‘fatigue’ under ‘other items’ and not as ‘feeling sluggish’, for this study the item ‘feeling sluggish’ was replaced by the better understood concept of ‘fatigue’ ( De Bom et al., 2008 ). We confirmed good internal consistency in our study (Cronbach’sα = 0.81).

During the first week of the chemotherapy cycle, patients completed the symptom checklist (immediate self-report) each day. To prevent the checklists from being completed too late, patients were offered to be reminded by telephone, e-mail or text message twice during the 7 days, once to complete their checklists and once to return the 7 checklists. During the hospital visit for the next chemotherapy administration, one of the researchers distributed an eighth checklist to the patients. Patients were now asked through both verbal instructions by the researchers and written instruction on the checklist to report the symptoms and side effects they had experienced during the first week of their chemotherapy cycle. To prevent response bias, when they were inducted into the study, patients were not informed that they would also be asked to complete a delayed self-report during their next visit. They were only told that a short evaluation would take place on their next hospital visit.


For the purpose of statistical analysis, the 7 data of the immediate self-report per symptom and per patient were transformed into 1 score. Since further therapeutic policy and supportive care are to be based on the highest score or the symptom severity at the worst day(s), the maximum score was considered to be the most relevant clinically. This is confirmed by guidelines from the U.S. National Cancer Institute’s CTCAE, which recommend that healthcare professionals report the worst severity of symptoms and side effects experienced by a patient within a chemotherapy cycle or since the last treatment (Basch et al, 2006 and National Cancer Institute, 2008). Indeed, and certainly for those symptoms of shorter duration, like chemotherapy-induced nausea and vomiting, caregivers need insight in the symptom burden on the days these symptoms occurred and mainly in how bad it was at that time, rather than an insight based on the average of those bad days and the days these symptoms didn’t occur. For this reason, the maximum score at the immediate self-report was used in all analyses presented. However, since we hypothesize that patients do not recall their worst symptom burden when asked about it at the next hospital visit but tend to tone down or average out between good and bad days, the analyses looking for correlation and difference between the immediate and delayed self-reported symptom severity were also conducted based on the median score of the immediate self-report.

Descriptive analyses were conducted for patient characteristics and symptom burden (number and grade/severity) of the study sample. Paired analyses were conducted to compare the number of symptoms and side effects per patient as well as the reported severity of the symptoms and side effects between the immediate and the delayed self-report. Wilcoxon signed-rank tests were carried out to compare the number of symptoms and side effects (grade ≥1) and the number of severe symptoms and side effects (grade ≥3) per patient. Spearman correlations were calculated to study the correlation between the grade reported at the late self-report and respectively the median and maximum grade reported at the immediate self-report. Exact marginal homogeneity tests were conducted to study a possible difference in reported grade and severity of symptoms and side effects. All data were analyzed using SPSS version 15.0 (SPSS, Inc., Chicago, IL).


Descriptive analysis of immediate- and delayed-reported symptom burden

Frequencies of symptoms at the immediate and delayed self-report are reported in Table 2 and reveal that profiles of symptom burden differ between the immediate and the delayed self-report.

Table 2 Occurrence and severity of symptoms and side effects. a

    Maximum score at the immediate self-report (%(n)) Delayed self-report (%(n))
Taste change Occurrence 81.3% (113) 73.5% (100)
Severity 30.2% (42) 20.6% (28)
Loss of appetite Occurrence 81.4% (114) 63.0% (87)
Severity 25.7% (36) 10.9% (15)
Nausea Occurrence 72.9% (102) 58.7% (81)
Severity 32.4% (46) 10.1% (14)
Vomiting Occurrence 32.4% (45) 19.0% (26)
Severity 31.9% (45) 5.8% (8)
Weight loss Occurrence 53.6% (74) 36.2% (50)
Severity 10.7% (15) 2.9% (4)
Sore mouth Occurrence 51.8% (71) 37.4% (52)
Severity 8.6% (12) 10.1% (14)
Cough Occurrence 35.1% (47) 23.9% (33)
Severity 13.9% (19) 2.2% (3)
Sore throat Occurrence 41.2% (56) 24.6% (33)
Severity 6.6% (9) 1.5% (2)
Difficulty swallowing Occurrence 42.5% (57) 29.2% (40)
Severity 8.8% (12) 4.4% (6)
Jaw pain Occurrence 29.1% (39) 15.2% (21)
Severity 10.3% (14) 1.4% (2)
Shortness of breath Occurrence 55.1% (75) 40.0% (54)
Severity 5.1% (7) 5.2% (7)
Numbness in fingers and/or toes Occurrence 55.8% (77) 46.0% (63)
Severity 10.1% (14) 11.7% (16)
Fatigue Occurrence 97.2% (138) 63.8% (88)
Severity 22.5% (32) 15.9% (22)
Depression Occurrence 35.8% (49) 24.8% (34)
Severity 45.8% (65) 2.9% (4)
Difficulty concentrating Occurrence 59.6% (81) 46.0% (63)
Severity 5.8% (8) 7.3% (10)
Fever Occurrence 15.6% (21) 5.8% (8)
Severity 10.9% (15) 2.2% (3)
Bruising Occurrence 9.0% (12) 9.5% (13)
Severity 2.2% (3) 0.7% (1)
Bleeding Occurrence 12.9% (17) 8.1% (11)
Severity 1.5% (2) 2.2% (3)
Hair loss Occurrence 60.3% (82) 55.0% (72)
Severity 4.3% (6) 21.4% (28)
Skin changes Occurrence 47.1% (64) 35.8% (49)
Severity 30.0% (42) 2.9% (4)
Soreness in vein receiving chemotherapy Occurrence 13.5% (18) 8.1% (11)
Severity 9.6% (13) 0.0% (0)
Difficulty sleeping Occurrence 73.6% (103) 47.1% (64)
Severity 34.3% (48) 10.3% (14)
Pain Occurrence 56.2% (77) 32.8% (44)
Severity 17.5% (24) 8.2% (11)
Decreased interest in sexual activity Occurrence 65.9% (87) 53.4% (62)
Severity 39.4% (52) 22.4% (26)
Constipation Occurrence 58.2% (78) 41.8% (51)
Severity 21.6% (29) 11.5% (14)

a Frequencies are calculated after dichotomization of the variables (occurrence: grade 1, 2, 3, and 4; severity: grade 3 and 4).

On the immediate self-report, the five most frequently reported symptoms (grade >1) were fatigue (97.2%); loss of appetite (81.4%); taste change (81.4%); nausea (72.9%) and hair loss (60.3%). Except for hair loss, these symptoms were also the most reported on the delayed self-report, but in different order and at a lower frequencies ( Table 2 ). Decreased interest in sexual activity appeared as the fifth most reported symptom at the delayed self-report.

On the immediate self-report, the most frequently reported severe symptoms (grade ≥3) were fatigue (45.8%); decreased interest in sexual activity (39.4%); difficulty of sleeping (34.3%); loss of appetite (32.9%); and nausea (32.1%). On the delayed self-report, a much lower proportion of patients reported fatigue (15.6%) and decreased interest in sexual activity (22.4%) as a severe symptom. Hair loss; taste change; and numbness in fingers and/or toes appeared in the top 5 of most reported severe symptoms at the delayed self-report, again at lower frequencies than at the immediate self-report.

Immediate-reported versus delayed-reported symptom burden

Number of reported symptoms

On the immediate self-report, the mean number (SD) of reported symptoms and side effects per patient was 11.9 (4.2), whereas on the delayed self-report, this value was 8.5 (4.4) (p < 0.05). For severe symptoms and side effects (grade ≥ 3), the mean number (SD) reported was 4.0 (3.7) on the immediate self-report and 1.8 (2.4) on the delayed self-report (p < 0.05).

Severity of reported symptoms

Spearman rank correlation coefficients ( Table 3 ) between the immediate-reported median score and the delayed self-report varied between 0.183 (soreness in vein receiving chemotherapy) and 0.824 (bleeding). Correlation coefficients between the immediate-reported maximum score and the delayed self-report ranged from 0.455 (bruising) to 0.785 (taste change).

Table 3 Spearman rank correlation coefficients, based on both the median and the maximum score at the immediate self-report.

  rS based on the median score at the immediate self-report rS based on the maximum score at the immediate self-report
Taste change 0.774 0.785
Loss of appetite 0.683 0.686
Nausea 0.709 0.766
Vomiting 0.507 0.740
Weight loss 0.623 0.493
Sore mouth 0.627 0.675
Cough 0.657 0.644
Sore throat 0.717 0.652
Difficulty swallowing 0.605 0.708
Jaw pain 0.637 0.597
Shortness of breath 0.775 0.662
Numbness in fingers and/or toes 0.799 0.762
Fatigue 0.552 0.509
Depression 0.721 0.753
Difficulty concentrating 0.685 0.735
Fever 0.324 0.619
Bruising 0.320 0.455
Bleeding 0.824 0.666
Hair loss 0.647 0.627
Skin changes 0.708 0.659
Soreness in vein receiving chemotherapy 0.183 0.489
Difficulty sleeping 0.736 0.726
Pain 0.662 0.564
Decreased interest in sexual activity 0.668 0.655
Constipation 0.643 0.643

Results on the exact marginal homogeneity tests ( Table 4 ) reveal that significant differences between the immediate- and delayed-reported symptom severity are very common. When comparing the delayed-reported score with the maximum score at the immediate self-report, statistical significant differences occur in 22 out of 25 symptoms of the TRSC (p < 0.05). In all these cases the delayed-reported score is significantly lower than the immediate-reported maximum grade. When comparing with the median score at the immediate self-report, statistical significant differences occur in only 10 out of 25 symptoms (p < 0.05). In 9 of these 10 symptoms the delayed-reported score is significantly higher than the immediate-reported median score. Only in fatigue, the significant difference goes in the opposite direction with the delayed-reported grade being significantly lower than the immediate-reported median.

Table 4 Results on exact marginal homogeneity tests (EMHT), based on both the median and the maximum score at the immediate self-report.

  Median score at the immediate self-report Maximum score at the immediate self-report Delayed self-report
Mean (SD) EMHT Mean (SD) EMHT Mean (SD)
Taste change 1.23 (1.17) p < 0.05 1.81 (1.28) p < 0.05 1.43 (1.67)
Loss of appetite 1.14 (1.10) 1.87 (1.28) p < 0.05 1.19 (1.15)
Nausea 0.83 (0.95) p < 0.05 1.66 (1.35) p < 0.05 1.04 (1.09)
Vomiting 0.15 (0.55) p < 0.05 0.69 (1.20) p < 0.05 0.37 (0.88)
Weight loss 0.28 (0.61) p < 0.05 0.86 (1.01) p < 0.05 0.59 (0.87)
Sore mouth 0.53 (0.96) p < 0.05 0.99 (1.17) p < 0.05 0.74 (1.13)
Cough 0.31 (0.68) 0.60 (0.97) p < 0.05 0.33 (0.66)
Sore throat 0.31 (0.71) 0.74 (1.07) p < 0.05 0.37 (0.72)
Difficulty swallowing 0.32 (0.67) p < 0.05 0.77 (1.10) p < 0.05 0.46 (0.84)
Jaw pain 0.19 (0.56) 0.49 (0.90) p < 0.05 0.24 (0.62)
Shortness of breath 0.46 (0.69) p < 0.05 0.96 (1.06) p < 0.05 0.66 (0.94)
Numbness in fingers and/or toes 0.85 (1.14) 1.27 (1.35) p < 0.05 0.91 (1.17)
Fatigue 1.72 (1.08) p < 0.05 2.39 (0.99) p < 0.05 1.22 (1.16)
Depression 0.35 (0.77) 0.66 (1.02) p < 0.05 0.42 (0.82)
Difficulty concentrating 0.63 (0.90) 1.07 (1.12) 1.01 (2.92)
Fever 0.04 (0.27) 0.24 (0.65) p < 0.05 0.12 (0.56)
Bruising 0.06 (0.35) 0.14 (0.54) 0.14 (0.47)
Bleeding 0.10 (0.38) 0.23 (0.65) 0.18 (0.67)
Hair loss 1.07 (1.38) 1.51 (1.51) p < 0.05 1.22 (1.40)
Skin changes 0.43 (0.72) p < 0.05 0.79 (1.00) p < 0.05 0.53 (0.81)
Soreness in vein receiving chemotherapy 0.08 (0.35) 0.26 (0.71) p < 0.05 0.11 (0.40)
Difficulty sleeping 0.90 (1.04) 1.71 (1.37) p < 0.05 0.87 (1.09)
Pain 0.61 (0.94) 1.22 (1.33) p < 0.05 0.63 (1.03)
Decreased interest in sexual activity 1.51 (1.54) 1.86 (1.58) p < 0.05 1.28 (1.43)
Constipation 0.62 (0.95) p < 0.05 1.27 (1.37) p < 0.05 0.83 (1.17)

Fig. 1 shows the frequencies of agreement and disagreement in grades between the immediate-reported maximum score and the delayed-reported grade of the ten most prevalent symptoms with significant difference. Frequencies of agreement varied between 25.4% (fatigue) and 63.8% (hair loss). Generally, grade differences tended to occur much more frequently in the left portions of Fig. 1 , indicating that in general patients gave these symptoms a lower grade on the delayed self-report than on the immediate self-report. A one-point lower grade at the delayed self-report occurred in 13.8%–37.0% of the cases. One-point grade differences in favor of the delayed self-report occurred only in 1.4–7.9% of the cases. Grade differences of two points or more resulted only rarely in a higher grade at the delayed self-report. However, two-point lower grades at the delayed self-report occurred in all ten symptoms, in 3.9% (hair loss) to 24.6% (fatigue) of the cases. In 0.8%–9.4% of the cases, three-point lower grades at the delayed self-report were found. In 8 of these 10 symptoms, 0.7–3.9% of the patients reported a four-point lower grade at the delayed self-report than their maximum immediate-reported score.


Fig. 1 Agreement and disagreement between immediate and delayed self-report.


The results of this study have clear-cut implications for several important issues in clinical practice. Patients had difficulty in accurately recalling symptoms that were experienced earlier in their chemotherapy treatment. Indeed, our patients reported fewer symptoms and fewer severe symptoms in the delayed self-report than on the immediate self-report. In the delayed self-report, they also reported significantly lower grades and severity than the maximum grade reported at the immediate self-report, and this for 22 symptoms out of the 25 questioned. For 9 symptoms the delayed-reported grade was significantly higher than the immediate-reported median grade. However, for fatigue, the delayed-reported grade was even significantly lower than the immediate-reported median grade.

Besides statistical significance, the study results should also be looked at from the perspective of clinical relevance. The discrepancy in the number of symptoms reported immediately and those reported later on is both statistically and clinically relevant, since symptoms not mentioned are likely to be left untreated ( Stromgren et al., 2001 ). The delay in symptom report resulted in a loss of approximately 3–4 symptoms per patients, excluding any action to support or relieve this symptom burden. Moreover, about 2 symptoms per patients were reported as severe at the immediate report but not at the delayed report. This means that therapeutic decisions or supportive care reserved to severe symptoms or grades 3 or higher, will fail to reach all patients needing them. The study also examined the qualitative discrepancy in grades reported immediately and delayed. For 22 out 25 symptoms patients reported significantly lower grades than the immediate-reported maximum grade, confirming our hypothesis that patients tend to tone down and that delayed symptom report fails to capture the actual experienced severity of symptoms. One-point grade differences are, as depicted in Fig. 1 , the most prevalent, though larger grade differences do occur at not negligible prevalence. It is difficult to determine whether a one-point grade difference is clinically relevant, since it depends on each symptom. Basch et al. (2006) stated that a shift in grade from 2 or less to 3 or more could already result in a clinically meaningful change in management. Given that any grade difference can be clinically relevant, Fig. 1 indicates the importance of the discrepancies between immediate-reported and delayed-reported maximum grades found in this study. On the other hand, the assumption that patients average out between good and bad days was only partly supported. For 15 out of 25 symptoms, no statistical significant differences were found between the delayed self-report and the immediate-reported median score, suggesting that for these symptoms patients indeed average out at the delayed symptom report. For 9 symptoms however, the delayed-reported grade was significantly higher than the immediate-reported median. Noteworthy is that fatigue was recalled significantly less severe at the delayed self-report than the immediate-reported median grade, suggesting a serious risk of this particular symptom to be minimized. Moreover, correlation coefficients between delayed-reported grade of fatigue and both immediate-reported median and maximum score were low (rS = 0.55 and 0.51 respectively).

Youngblood et al. (1994) and Homsi et al. (2006) recommended the use of written symptom self-reports, because they found a significant difference in the number of symptoms reported orally versus that reported in writing. The present study found that symptoms tend to be underreported both quantitatively and qualitatively on delayed written self-reports. Therefore, immediate self-reporting of chemotherapy symptoms and side effects are recommended.

Further study is needed to examine whether focusing delayed self-reports on the worst symptom burden, thus asking patients to report their worst experienced symptom severity, provides more accurate insight.

There are several options that can facilitate written and immediate patient self-report of chemotherapy symptoms in clinical research and practice. One option is patient diaries. Another option is interactive media that implement the Internet or mobile phone technology. Indeed, modern media are more and more used, studied, and appreciated as tools for real-time symptom tracking and reporting (Basch et al, 2005, Basch et al, 2007, Kearney et al, 2006, and Weaver et al, 2007). Such interactive systems not only allow accurate insight into a patient’s symptom burden but also permit timely and appropriate responses from clinicians and nurses (Basch et al, 2005 and Kearney et al, 2006). Logically, the earlier side effects are reported, the quicker they can be managed and the better they can be minimized ( Weaver et al., 2007 ).

Next, we address some methodological considerations, as these affect the interpretation of the study results. Firstly, the instrument used in the present study may not have been optimal. Remarkably, no less than 17 different manifestations of pain were reported under ‘other problems’, while the TRSC itself explicits 5 manifestations of pain among which ‘pain’ as a general symptom. Furthermore, 48 symptoms not associated with any item of the TRSC, were listed under the ‘other problems’ item. Most prevalent were diarrhea (n = 21), xerostomia or dry mouth (n = 9) and watery eyes (n = 7).

Secondly, the study itself may possibly introduce bias into the phenomenon studied. Data were obtained by having the patients complete a symptom checklist, which in itself could have put emphasis on the symptoms the patients experienced. To reduce possible bias caused by the study itself, we limited the duration of the immediate self-reports to the first week of the chemotherapy cycle. However, since some symptoms arise later on, limiting the reporting period to 7 days also includes the risk of missing some symptoms. Moreover, with regard to reporting symptoms on the delayed self-report, it might have been difficult for the patients to consider and recall only the symptoms that they had experienced during the first week of the chemotherapy cycle.

Since examining the reasons why immediate and delayed self-report data differed was beyond the scope of this study, the possible underlying mechanisms remain unclear. Response shift has been described as a possible mechanism attenuating estimates of side effects of treatment. The term response shift was introduced by Howard et al. (1979) and defined in terms of changes in internal standards of measurement. Several methods are available to study a possible response shift in patient self-report (Jansen et al, 2000, Oort et al, 2009, and Visser et al, 2005). None of them were applied in our study. However, the retrospective or delayed self-report of symptoms is strongly reminiscent to the then-test conducted in research studying response shift. Using the same questionnaire as the pre-test, the then-test asks patients to provide a new judgment about their quality of life at the time they completed the pre-test. As was the case in this study, patients are asked to provide a new judgment of their quality of life at the time rather than recalling their responses at the first questionnaire. Thus far response shift has mainly been studied for global quality of life before and after diagnosis or treatment, and less at the level of individual symptoms during treatment. Other mechanisms possibly adding to the explanation of underreporting of actual problems, are self-presentation and recall bias. Self-presentation results from the need to be positive and to distance oneself from a stereotypical cancer patient and was described by Westerman et al. (2007) as a partial explanation of counter-intuitive findings in the self-report of fatigue by small-cell lung cancer patients. In recall bias, distorted retrospective scores are the consequence of patients remembering in particular the most intense or the last experience of an episode ( Jansen et al., 2000 ). None of these mechanisms can be ruled out or confirmed as explanations of this study’s findings.

Another possible influence to the research question and to the way patients recall symptoms, not investigated in this study, lies within the concept of the symptom experience. This symptom experience is the perception of not only frequency and severity/intensity, as measured in this study, but also of symptom distress and meaning (Armstrong, 2003 and Gapstur, 2007). The latter haven’t been questioned in the study. It remains therefore unclear whether the distress resulting from experiencing symptoms, or the meaning or importance of symptoms to patients influence patients’ delayed self-report.

While explaining the difference between immediate and delayed self-report of symptoms during chemotherapy merits further study, it is important to stress that evidence based clinical practice of symptom report during chemotherapy shouldn’t wait for further explanatory research.


The results of this study confirm the hypothesis that delayed symptom self-report of chemotherapy side effects provides a distorted, weaker overview of the actual symptoms and symptom severity experienced by patients. Respondents reported fewer symptoms in the delayed self-report than in the immediate self-report. In the delayed self-report respondents reported 22 out of 25 symptoms significantly less severe than their worst grade at the immediate self-report. For 15 out 25 symptoms patients seem to average out between good and bad days at the delayed symptom recall. In the other cases, except for fatigue, delayed-reported grades were significantly higher than the immediate-reported median grade. The results of this study justify the recommendation of immediate, written self-report tools as a method for accurately assessing symptoms during chemotherapy.


  • Armstrong, 2003 T.S. Armstrong. Symptoms experience: a concept analysis. Oncology Nursing Forum. 2003;30:601-606 Crossref
  • Basch et al., 2005 E. Basch, D. Artz, D. Dulko, K. Scher, P. Sabbatini, M. Hensley, N. Mitra, J. Speakman, M. McCabe, D. Schrag. Patient online self-reporting of toxicity symptoms during chemotherapy. Journal of Clinical Oncology. 2005;23:3552-3561 Crossref
  • Basch et al., 2007 E. Basch, D. Artz, A. Iasonos, J. Speakman, K. Shannon, K. Lin, C. Pun, H. Yong, P. Fearn, A. Barz, H.I. Scher, M. McCabe, D. Schrag. Evaluation of an online platform for cancer patient self-reporting of chemotherapy toxicities. Journal of the American Medical Informatics Association. 2007;14:264-268 Crossref
  • Basch et al., 2006 E. Basch, A. Iasonos, T. McDonough, A. Barz, A. Culkin, M.G. Kris, H.I. Scher, D. Schrag. Patient versus clinician symptom reporting using the National Cancer Institute Common Terminology Criteria for Adverse Events: results of a questionnaire-based study. The Lancet Oncology. 2006;7:903-909 Crossref
  • Braud et al., 2003 A.C. Braud, D. Genre, C. Leto, V. Nemer, J.F. Cailhol, G. Macquart-Moulin, D. Maraninchi, P. Viens. Nurses’ repeat measurement of chemotherapy symptoms: feasibility, resulting information, patient satisfaction. Cancer Nursing. 2003;26:468-475
  • Butt et al., 2008 Z. Butt, L.I. Wagner, J.L. Beaumont, J.A. Paice, A.H. Peterman, D. Shevrin, J.H. Von Roenn, G. Carro, J.L. Straus, J.C. Muir, D. Cella. Use of a single-item screening tool to detect clinically significant fatigue, pain, distress, and anorexia in ambulatory cancer practice. Journal of Pain and Symptom Management. 2008;35:20-30 Crossref
  • De Bom et al., 2008 De Bom, A., Coolbrandt A., Vanhove, E., Wildiers, H., Milisen, K., 2008. Immediate versus delayed self-report of side effects during chemotherapy: a pilot study. Unpublished Dissertation, Nursing Sciences, Catholic University, Leuven.
  • Fromme et al., 2004 E.K. Fromme, K.M. Eilers, M. Mori, Y.C. Hsieh, T.M. Beer. How accurate is clinician reporting of chemotherapy adverse effects? A comparison with patient-reported symptoms from the quality-of-life questionnaire C30. Journal of Clinical Oncology. 2004;22:3485-3490 Crossref
  • Gapstur, 2007 R.L. Gapstur. Symptom burden: a concept analysis and implications to oncology nurses. Oncology Nursing Forum. 2007;34:673-680 Crossref
  • Homsi et al., 2006 J. Homsi, D. Walsh, N. Rivera, L.A. Rybicki, K.A. Nelson, S.B. Legrand, M. Davis, M. Naughton, D. Gvozdjan, H. Pham. Symptom evaluation in palliative medicine: patient report vs systematic assessment. Supportive Care in Cancer. 2006;14:444-453 Crossref
  • Howard et al., 1979 G.S. Howard, K.M. Ralph, N.A. Gulanick, S.E. Maxwell, D.W. Nance, S.K. Gerber. Internal invalidity in pretest-posttest self-report evaluations and re-evaluation of retrospective pretests. Applied Psychological Measurement. 1979;3:1-23 Crossref
  • Jansen et al., 2000 S.J.T. Jansen, A.M. Stiggelbout, M.A. Nooij, E.M. Noordijk, J. Kievit. Response shift in quality of life measurement in early-stage breast cancer patients undergoing radiotherapy. Quality of Life Research. 2000;9:603-615 Crossref
  • Kearney et al., 2006 N. Kearney, L. Kidd, M. Miller, M. Sage, J. Khorrami, M. McGee, J. Cassidy, K. Niven, P. Gray. Utilising handheld computers to monitor and support patients receiving chemotherapy: results of a UK-based feasibility study. Supportive Care in Cancer. 2006;14:742-752 Crossref
  • Lobchuk and Degner, 2002a M.M. Lobchuk, L.F. Degner. Symptom experiences: perceptual accuracy between advanced-stage cancer patients and family caregivers in the home care setting. Journal of Clinical Oncology. 2002;20:3495-3507 Crossref
  • Lobchuk and Degner, 2002b M.M. Lobchuk, L.F. Degner. Patients with cancer and next-of-kin response comparability on physical and psychological symptom well-being: trends and measurement issues. Cancer Nursing. 2002;25:358-374 Crossref
  • National Cancer Institute, 2008 National Cancer Institute. CTC v2.0 and Common Terminology Criteria for Adverse Events v3.0 (CTCAE). (, 2008) Retrieved March 31, 2009, from
  • Oort et al., 2009 F.J. Oort, M.R.M. Visser, M.A.G. Sprangers. Formal definitions of measurement bias and explanation bias clarify measurement and conceptual perspectives on response shift. Journal of Clinical Epidemiology. 2009;62:1126-1137 Crossref
  • Stromgren et al., 2001 A.S. Stromgren, M. Groenvold, L. Pedersen, A.K. Olsen, M. Spile, P. Sjogren. Does the medical record cover the symptoms experienced by cancer patients receiving palliative care? A comparison of the record and patient self-rating. Journal of Pain and Symptom Management. 2001;21:189-196 Crossref
  • Visser et al., 2005 M.R.M. Visser, F.J. Oort, M.A.G. Sprangers. Methods to detect response shift in quality of life data: a convergent validity study. Quality of Life Research. 2005;14:629-639 Crossref
  • Weaver et al., 2007 A. Weaver, A.M. Young, J. Rowntree, N. Townsend, S. Pearson, J. Smith, O. Gibson, W. Cobern, M. Larsen, L. Tarassenko. Application of mobile phone technology for managing chemotherapy-associated side-effects. Annals of Oncology. 2007;18:1887-1892 Crossref
  • Westerman et al., 2007 M.J. Westerman, A. The, M.A.G. Sprangers, H.J.M. Groen, G. van der Wal, T. Hak. Small-cell lung cancer patients are just ‘a little bit’ tired: response shift and self-presentation in the measurement of fatigue. Quality of Life Research. 2007;16:853-861 Crossref
  • Williams et al., 2001 P.D. Williams, K.A. Ducey, A.M. Sears, A.R. Williams, S.E. Tobin-Rumelhart, P. Bunde. Treatment type and symptom severity among oncology patients by self-report. International Journal of Nursing Studies. 2001;38:359-367 Crossref
  • Williams et al., 2006 P.D. Williams, U. Piamjariyakul, K. Ducey, J. Badura, K.D. Boltz, K. Olberding, A. Wingate, A.R. Williams. Cancer treatment, symptom monitoring, and self-care in adults: pilot study. Cancer Nursing. 2006;29:347-355 Crossref
  • Youngblood et al., 1994 M. Youngblood, P.D. Williams, H. Eyles, J. Waring, S. Runyon. A comparison of two methods of assessing cancer therapy-related symptoms. Cancer Nursing. 1994;17:37-44


a Department of Oncology/Hematology, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium

b Nursing Competence Center, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium

c Center for Health Services and Nursing Research, Katholieke Universiteit Leuven, Kapucijnenvoer 35/4, 3000 Leuven, Belgium

d Department of Geriatric Medicine, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium

e Department of Medical Oncology, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium

Corresponding author. Tel.: +32 16 341314; fax: +32 16 346772.

Editors’ comment: Dr Ricardo Caponero

Some electronic devices allow immediate patient reporting of any adverse event. We suppose that patients forget side effects as time goes by. Is it true? Annemarie Coolbrandt and coworkers studied immediate versus delayed self-reporting of symptoms and conclude that delayed self-reporting of chemotherapy side effects is not an appropriate measure of actual symptoms and side effects experienced by patients, and that fatigue is at particular risk to be minimized in delayed self-reporting. As we suspected, prompt reporting is more accurate.

Search this site

Featured videos

Free access to ebooks on cachexia and CINV


Prevention of nausea and vomiting in
adult cancer patients receiving tumour-
directed therapy – 2017 update –

Cancer cachexia: mechanisms and progress in treatment


Stay up-to-date with our monthly e-alert

If you want to regularly receive information on what is happening in Quality of Life in Oncology research sign up to our e-alert.

Subscribe »

QOL (Quality of Life) newsletter e-alert

Quality of Life promotional video

Made possible by an educational grant from Helsinn

Helsinn does not have any influence on the content and all items are subject to independent peer and editorial review

Society Partners

European Cancer Organisation Logo