Electronic self-report assessment for cancer (ESRA-C): Results of a randomized clinical trial
Reviewer: Christine Hill, MD
Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 3, 2008
Presenter: D.L. Berry Presenter's Affiliation: University of Washington, Seattle, WA Type of Session: Scientific
As the healthcare milieu continuously evolves, healthcare providers are faced with the difficulties of shrinking resources, which may ultimately impact upon the amount of time spent with patients during individual clinic visits.
The patient experience, however, remains essential to the development of plans for medical care and treatment.
Assessment of common cancer symptoms and quality of life (QOL) is very important to overall patient care; however, addressing the many possible symptoms that patients may encounter can be difficult during a visit limited to 10 – 20 minutes; additionally, patients may be uncomfortable addressing concerns that are not initially raised by the physician.
This may be particularly true with regards to symptoms having to do with psychological or sexual well-being (Woodhouse, 2008).
Electronic self-report methods have and are being developed, with several potential benefits:
Systems can be made available on several different types of devices, including office-computers, personal digital assistants (PDA), telephones, or the internet in order to be readily accessible to patients.
Electronic systems provide a customized, confidential, and private method for gathering information.
Use of these systems within the healthcare system can allow much more ready access to data required for research, and can eliminate the steps of data abstraction that can be quite time-consuming and cumbersome.
Despite these potential benefits, little data regarding actual benefits to patients have been demonstrated using electronic reporting methods.
This randomized, controlled, trial was designed to investigate the use of an electronic self-report assessment for cancer (ESRA-C), particularly with regards to patient/ physician communication, and the overall impact of this intervention on clinical care.
Materials and Methods
This randomized, controlled trial enrolled patients from two urban cancer centers in the Seattle, WA area.
Patients were recruited from ambulatory centers (both medical oncology and radiation oncology offices). Patients undergoing stem cell transplant (SCT) were also included.
Patients with all types and stages of cancer were included.
Enrollment was limited to two patients per provider per month in order to attempt to avoid practice bias.
All patients were asked to complete the ESRA-C on two different occasions – first before initiation of treatment, and then again either at discharge from the hospital (for patients undergoing SCT), or 6 weeks after completing ESRA-C in the ambulatory care setting.
The ESRA-C was completed on touch screen computers, and contained questions regarding common cancer symptoms and QOL measures.
Quality of life was assessed via the European Organization for Research and Treatment of Cancer Core Quality of Life Questionnaire (EORTC QLQ-30), which was designed specifically for QOL determination in cancer patients (Aaronson, 1993).
Depression was assessed using a nine-item depression scale of the Patient Health Questionnaire (PHQ9).
Significance of symptoms to patients was evaluated using the Subjective Significance Scale (SSQ).
Impact on sexual function and pain intensity were posed as individual items.
The ESRA-C is designed to pose one question per screen. The computer screen darkens after a question is answered; however, patients can skip questions and change prior answers when using this tool.
All patients enrolled on this study completed two ESRA-C assessments. For patients in the intervention group, a graphical, color summary of responses was given to clinicians prior to their seeing the patient. For patients in the control group, no graphical summary was provided to the treatment team.
The graphical summary consisted of red and green bars reflecting responses regarding single symptoms, and effects on QOL based on the EORTC QLQ-30 subscales.
Office visits on the days on which ESRA-C was completed were audio-recorded. They were abstracted by research coordinators blinded to whether patients were randomized to either the intervention or the control group.
Research coordinators noted times when clinicians raised issues that patients had also raised via ESRA-C. This was recorded as a binary response (discussed or not discussed) for each issue that patients had identified on the ESRA-C.
They also noted treatments for symptoms that were initiated.
Generalizing estimating equations (GEE) and logistic regression modeled the odds of discussion of 27 symptoms and QOL concerns by study arm.
Over the period of two years (spring 2005 – summer 2007), 1,104 patients were approached regarding enrollment on this trial.
Of these, 796 agreed and were consented, and 590 completed two ESRA-C assessments.
Over the same time period, 295 clinicians were approached, and 76 were ultimately involved in the trial.
Patient background characteristics did not differ significantly between the two arms, with the exception that a greater number of patients with low-income were present in the control arm (29% versus 22% in the intervention arm).
Approximately 50% of patients in each arm were being seen in a medical oncology clinic.
Approximately 30% were undergoing SCT.
The remaining 20% were being seen in radiation oncology clinics.
When results were assessed in a binary fashion, the probability of an issue being discussed was found to depend on whether it was reported by the patient as moderate or severe versus mild. This result was true in both the intervention and control groups.
The probability of several issues being discussed also depended on the availability of the graphical ESRA-C summary (p < 0.015).
On logistic regression, the odds of discussion of issues such as sexual interest and activity, emotional function, and cognitive function were increased significantly when the graphical summary was provided, with odds ratios of 2.9, 1.7, and 1.8, respectively.
Within patients being seen in a radiation oncology clinic, the odds of discussing appetite were significantly higher in the intervention group, with an odds ratio of 2.5.
Concerns about physical appearance were discussed less often in the intervention SCT group, with an odds ratio of 0.15.
Concerns regarding sexual interest and activity were raised in the ESRA-C by 120 patients in the intervention group, and 122 in the control group. These issues were addressed by clinicians in 27 of 120 cases when the graphical summary was available, and 10 of 122 cases when it was not. These differences were statistically significant (p < 0.05). Of note, of the 10 patients with whom sexual issues were discussed in the absence of the graphical summary, all were undergoing treatment for gynecologic or genitourinary cancers.
Concerns regarding depression were raised by 62 patients in the intervention group, and 54 in the control group. Depression was discussed in the cases of 60 of 64 patients for whom graphical summary was provided, and 40 of 54 patients for whom it was not. Again, these differences were statistically significant (p < 0.05).
These differences did not appear to exist with regard to either nausea or pain, both of which were almost always discussed either with or without the graphical summary.
The authors conclude that the ESRA-C represents the first electronic self-reporting system to be evaluated in the setting of a randomized, controlled trial in the United States.
They conclude that it is an effective tool for promoting discussion and increasing patient-healthcare provider communication, and was helpful in focusing the clinic visit on relevant patient concerns.
This trial represents a well-designed and well-executed assessment of the role of electronic self-reporting of symptoms in cancer patients.
The questions addressed in this trial are extremely pertinent and important as the field of medicine evolves to be paperless and electronically based.
Certainly, systems such as the ESRA-C are expected to be implemented more and more frequently within the clinical setting. In addition to the benefits raised by the authors, electronic self-reporting systems can be interfaced with electronic medical records, allowing frequent and efficient placement of symptoms and concerns in patient charts.
Interestingly, although the ESRA-C increased physician discussion of depression, sexual function, and cognitive issues, actual discussion of sexual issues remained sparse, with sexual function being addressed for only 27 patients of 120 reporting sexual problems, even when the graphical summary was available; certainly, this individual issue needs to continue to be addressed in order to increase physician comfort and willingness to discuss it.
Although electronic self-reporting is very appealing in theory, the authors addressed the important questions of the contribution of these systems to actual patient care.
Their findings that electronic self-reporting increases communication between patients and healthcare providers are interesting and important, and have been validated by other groups (Velikova, 2004; Rosenbloom, 2007).
In studies that have assessed the contribution of this improved communication on QOL, medical decision-making, and well-being, however, no impact has been seen (Velikova, 2004; Rosenbloom, 2007; McLachlan, 2001; Detmar, 2002).
These findings certainly raise concern that benefits seen in communication with electronic-self reporting are not being translated into improved medical care or improved quality of life. These issues were not addressed in this study, but would be interesting to address in the future.
This study is indeed somewhat limited by the fact it did not assess outcomes, as well as the lack of assessment of symptoms discussed at baseline for each patient. Certainly, one would expect the probability of discussion of an issue to increase if it had been discussed at a previous visit; this likelihood could confound the data presented here, and was not accounted for in this study.
Additionally, a relatively small number (76) of the 296 clinicians originally approached about this study actually participated in it. This raises the concern that physicians participating in this trial were enthusiastic about use of the ESRA-C system, and that their results may not be applicable to the general clinician population.
Having said this, this study represents a well-designed assessment of a type of system that will in all likelihood become a widespread part of clinical patient care in the future. Continued assessment of the impact of systems such as ESRA-C on clinical outcomes, as well as techniques to increase this impact, is essential.
Nov 2, 2010 - Most recent oncology randomized controlled trials evaluate drugs that are available "off-protocol therapy" in the United States, and this can adversely impact trial enrollment, according to a study published online Oct. 25 in the Journal of Clinical Oncology.