Abstract
Objective
To integrate multiple sources of clinical information with patient feedback to build evidence-based decision support model to facilitate treatment selection for patients suffering from vestibular schwannomas (VS).
Methods
This was a mixed methods study utilizing focus group and survey methodology to solicit feedback on factors important for making treatment decisions among patients. Two 90-minute focus groups were conducted by an experienced facilitator. Previously diagnosed VS patients were recruited by clinical investigators at the University of Pittsburgh Medical Center (UPMC). Classical content analysis was used for focus group data analysis. Providers were recruited from practices within the UPMC system and were surveyed using Delphi methods. This information can provide a basis for multi-criteria decision analysis (MCDA) framework to develop a treatment decision support system for patients with VS.
Results
Eight themes were derived from these data (focus group + surveys): doctor/health care system, side effects, effectiveness of treatment, anxiety, mortality, family/other people, quality of life, and post-operative symptoms. These data, as well as feedback from physicians were utilized in building a multi-criteria decision model.
Discussion
The study illustrated steps involved in the development of a decision support model that integrates evidence-based data and patient values to select treatment alternatives.
Conclusions
Studies focusing on the actual development of the decision support technology for this group of patients are needed, as decisions are highly multifactorial. Such tools have the potential to improve decision making for complex medical problems with alternate treatment pathways.
1
Introduction
A vestibular schwannoma (VS) or acoustic neuroma, is a benign, slow-growing tumor that arises from the 8th cranial nerve, which controls balance and hearing in the inner ear . VS accounts for roughly 8% of newly diagnosed intracranial neoplasms , with an incidence rate of approximately 10–15 cases per million per year in the United States . Mortality rates are estimated to be as low as 0.4% . VS tumors arise due to an over-production of Schwann cells and as the tumor grows within the narrow confines of the internal auditory canal and cerebellopontine angle, pressure is exerted on the vestibular and cochlear nerves. The primary symptoms are asymmetric hearing loss, tinnitus, and disequilibrium. Secondary symptoms consist of facial numbness and weakness (if tumor affects the 5th and 7th cranial nerves), changes in taste, and difficulty in swallowing and hoarseness if the lower cranial nerves are affected by larger tumors. Diagnosis of VS is made with radiographic imaging such as magnetic resonance imaging (MRI) scans.
The three main treatment options for VS are surgery, stereotactic radiosurgery (SRS)/stereotactic radiotherapy (SRT), and conservative management (CM). The optimal treatment is very difficult to establish, as each treatment has its possible advantages, complications, and limitations. Due to the anatomical location of VS, the morbidity and potential side effects of treatment on quality of life are profound with alterations in balance, communication, and appearance. Surgery is more likely to be recommended for large symptomatic tumors. For those with small tumors (<1.5 cm), surgery, SRS/SRT and CM are all options.
Comparison of treatments is confounded by the variability of patients, tumor size and location, which potentiates different morbidities associated with each treatment option. Health care providers are not unbiased and a recommendation for treatment differs among providers. To our knowledge, no decision support system has been developed to date to assist patients and/or providers with this difficult decision.
Multi-criteria decision analysis (MCDA) can provide an approach for building an effective patient–physician interaction for identifying the best treatment, individualized to the needs and values of VS patients. MCDA is a sub-discipline of operations research that explicitly considers multiple criteria in decision-making environments. It integrates medical information by defining criteria and metrics associated with treatment alternatives based on evidence-based data. In diagnoses where there is a lack of consensus among medical professionals in treatment efficacy or likelihood of comorbidities following treatment, the MCDA methods potentially provide an opportunity for patients to assign their personal value to the treatment outcomes. Implementing the MCDA process thus will not only provide information to the patient in a structured way, it will have the power to combine individual patient preferences with evidence-based data to derive the best option for the patient, making this technology potentially useful for all decisions at all levels in the health care system. MCDA tools have been used across multiple settings and diseases . Feasibility and implementation studies, like ours, have the potential to provide important information to help guide development, evaluation, and implementation of decision support tools and ultimately improve health care systems. Previous publications suggested that MCDA is essential for prioritizing treatment options in the context of high uncertainty , which may be the case when evidence-based medical information is incomplete, multiple management options exist, and multiple stakeholders are involved.
The goal of this paper is to systematically gather evidence from literature, patients, and providers to identify factors important for decision making in patients with VS and to use that evidence to develop a decision support tool for VS patients.
1.1
Treatments
Treatment for VS typically depends on the size and symptoms of the tumor, the rate of growth of the tumor, and patient characteristics such as age and medical comorbidities.
1.1.1
Surgery
The primary goal of surgical approaches is the complete removal of the tumor. Surgical approaches offer high rates of tumor excision, low rates of tumor recurrence, and preservation of hearing and facial nerve function for small tumors . Surgery is typically contraindicated in the elderly and patients with co-morbidities. Serious complications from surgical approaches include facial paresis, cerebrovascular accidents, cerebrospinal fluid leak, meningitis, and prolonged headaches after surgery . When compared to other treatment options, surgery does have higher rates of hearing loss and facial paralysis , and several studies have shown that roughly 50% of patients undergoing surgery report lower quality of life scores .
1.1.2
Stereotactic radiosurgery (SRS)/stereotactic radiotherapy (SRT)
The primary goal of radiosurgery and radiotherapy is tumor control. In both of these approaches, the tumor remains intact, but changes in volume and consistency of the tumor are also frequent. Tumor re-growth is more common in these approaches, as compared to surgery; therefore, if these approaches fail, surgical salvage is typically the treatment option of choice . Surgery may be more difficult in those situations with a greater risk of complications.
SRS can be delivered through several different systems and types of radiation. These include the Gamma Knife system, linear accelerator (LINAC)-based systems such as CyberKnife, and proton beam therapy. Gamma Knife, for example, delivers a high dose of radiation in one treatment. This approach has higher rates of hearing preservation and facial nerve function than surgical approaches and is far less invasive . Gamma Knife SRS is completed in one day and is relatively painless, patients return to normal activities faster, and the cost is less than surgery. However, this approach still carries with it a risk of hearing loss and facial paralysis, and should be used with caution when treating tumors compressing the brainstem. Complications from SRS include radiation damage to the surrounding cranial nerves and brainstem, the risk of developing radiation-induced cancers years after the treatment, and progression of tumor growth.
SRT delivers multiple small doses of radiation over time (fractionation). This approach has the highest rate of hearing preservation , high rates of tumor control , and reduced radiation necrosis to the brainstem and cranial nerves. The primary disadvantage to this option is the need for up to 30 treatments over 5 days per week for 6 weeks. As with SRS, there is a risk of developing radiation-induced cancer years after the treatment, and progression of tumor growth.
1.1.3
Conservative management (CM)
This is a “wait-and-scan” approach which can offer select patients an alternative to open surgery and/or exposure to radiation . Patients with stable tumors and few or no symptoms may not need an intervention but can be serially monitored by MRI to ensure there is no growth. In the event that tumor growth is detected and/or increased symptoms occur, active treatment with surgery or radiation can be undertaken. In one study by Fucci et al., 70% of tumors <20 mm did not grow and 4% regressed .
2
Materials and methods
Using mixed methods is the most appropriate strategy to obtain a multifaceted understanding of factors influencing patient decisions and define characteristics to be included in the VS decision support tool. The methodology included the following stages:
Stage I was a focus group comprised of patients previously diagnosed with VS, caregivers of VS patients, health care providers, and researchers. It consisted of 2 patients, 2 providers (physician and nurse both taking care of patients with VS), and 1 caregiver. One patient already underwent a surgery and the other was in the stage of observation. It was from these participants that we were able to determine the factors that were considered important and needed to be included in the decision pathway.
Stage II included 8 MD level providers (neurotologists, neurosurgeons, and radiosurgeons) who were surveyed using Delphi methods to determine and quantify the risks and benefits of each treatment alternative based on criteria identified in Stage I. One of the authors (CS) developed a list of experts treating patients with VS and invited them to participate; 8 physicians agreed to provide information. The Delphi survey 1 requested a list of potential complications for three key treatment options for VS treatment. Delphi survey 2 asked physicians to estimate percent risk of developing such complication for each treatment modality. We calculated a mean score for each item reviewed by physicians by summing all ratings reported for a single item. Survey 3 shared group results with all physicians along with supporting data based on literature review; physicians were invited to adjust their original scoring for any items. Consensus regarding risk levels was achieved after Delphi survey 3. The Delphi approach has been extensively used for achieving expert consensus in health care settings . In this case, use of Delphi technique had several key advantages including the ability to bring together feedback from a relatively large group of experts, ensuring confidentiality of responses, structured group communication process, and convenience.
Stage III involved the development of the multi-criteria decision model using data collected in stages I and II following a systematic decision framework : 1) generating alternative treatment options, 2) developing success criteria and metrics for evaluating alternatives, 3) building in ability to elicit value judgments on the importance of criteria and metric relevance, and 4) scoring the alternatives by applying value weights.
Stage IV was a second focus group consisting of 4 previously diagnosed VS patients (already treated for VS) and one caregiver, where we further explored the decision making process retrospectively through the administration of surveys, and tested the feasibility of implementing the decision pathway. While these patients already underwent surgical treatment for VS, these patients were previously diagnosed with small (<1.5 cm) tumors, and experienced a decision-making paradigm where more than one appropriate choice of treatment was available to them.
Both focus groups utilized an interview guide, an experienced facilitator, to conduct a 90-minute focus group. As a part of the focus group, we administered short surveys, which focused on identifying the factors important for decision-making. In addition, notes were recorded during the focus group discussion in order to capture any decision-making factors that may have been missed by the surveys. Classical content analysis was used for all focus group data analysis. Classical content analysis is the process of systematically reducing qualitative data to code and then counting the number of codes . For example, using the surveys and the notes taken during the focus group discussion, we identified all factors that played a role in the decision-making process. For each factor, we assigned a code and then counted all instances of a given code. Six VS patients for the focus groups were recruited by clinical investigators from UPMC practices. Participants that were unable to attend meetings in person completed online surveys via email interaction with the program coordinator.
Focus group discussion was centered on four major questions ( Table 1 ). The model presented in Fig. 1 represents data obtained for small tumors, as these are the tumors where patients are presented with a true decision paradigm about the kinds of choices available to them. Fig. 2 summarizes the key points of the research methods. This project received ethical approval as a quality improvement project (University of Pittsburgh protocol #0001413).
| 1) What factors influenced your decision to undergo surgical treatment (if applicable)? | |
| 2) Which people were important in your decision making? What assumptions were you making? | |
| 3) How satisfied were you with the treatment outcome and would you change anything if you could go back to your decision point? | |
| 4) How would you rate the potential usefulness of the decision tree in making your choices? |
2
Materials and methods
Using mixed methods is the most appropriate strategy to obtain a multifaceted understanding of factors influencing patient decisions and define characteristics to be included in the VS decision support tool. The methodology included the following stages:
Stage I was a focus group comprised of patients previously diagnosed with VS, caregivers of VS patients, health care providers, and researchers. It consisted of 2 patients, 2 providers (physician and nurse both taking care of patients with VS), and 1 caregiver. One patient already underwent a surgery and the other was in the stage of observation. It was from these participants that we were able to determine the factors that were considered important and needed to be included in the decision pathway.
Stage II included 8 MD level providers (neurotologists, neurosurgeons, and radiosurgeons) who were surveyed using Delphi methods to determine and quantify the risks and benefits of each treatment alternative based on criteria identified in Stage I. One of the authors (CS) developed a list of experts treating patients with VS and invited them to participate; 8 physicians agreed to provide information. The Delphi survey 1 requested a list of potential complications for three key treatment options for VS treatment. Delphi survey 2 asked physicians to estimate percent risk of developing such complication for each treatment modality. We calculated a mean score for each item reviewed by physicians by summing all ratings reported for a single item. Survey 3 shared group results with all physicians along with supporting data based on literature review; physicians were invited to adjust their original scoring for any items. Consensus regarding risk levels was achieved after Delphi survey 3. The Delphi approach has been extensively used for achieving expert consensus in health care settings . In this case, use of Delphi technique had several key advantages including the ability to bring together feedback from a relatively large group of experts, ensuring confidentiality of responses, structured group communication process, and convenience.
Stage III involved the development of the multi-criteria decision model using data collected in stages I and II following a systematic decision framework : 1) generating alternative treatment options, 2) developing success criteria and metrics for evaluating alternatives, 3) building in ability to elicit value judgments on the importance of criteria and metric relevance, and 4) scoring the alternatives by applying value weights.
Stage IV was a second focus group consisting of 4 previously diagnosed VS patients (already treated for VS) and one caregiver, where we further explored the decision making process retrospectively through the administration of surveys, and tested the feasibility of implementing the decision pathway. While these patients already underwent surgical treatment for VS, these patients were previously diagnosed with small (<1.5 cm) tumors, and experienced a decision-making paradigm where more than one appropriate choice of treatment was available to them.
Both focus groups utilized an interview guide, an experienced facilitator, to conduct a 90-minute focus group. As a part of the focus group, we administered short surveys, which focused on identifying the factors important for decision-making. In addition, notes were recorded during the focus group discussion in order to capture any decision-making factors that may have been missed by the surveys. Classical content analysis was used for all focus group data analysis. Classical content analysis is the process of systematically reducing qualitative data to code and then counting the number of codes . For example, using the surveys and the notes taken during the focus group discussion, we identified all factors that played a role in the decision-making process. For each factor, we assigned a code and then counted all instances of a given code. Six VS patients for the focus groups were recruited by clinical investigators from UPMC practices. Participants that were unable to attend meetings in person completed online surveys via email interaction with the program coordinator.
Focus group discussion was centered on four major questions ( Table 1 ). The model presented in Fig. 1 represents data obtained for small tumors, as these are the tumors where patients are presented with a true decision paradigm about the kinds of choices available to them. Fig. 2 summarizes the key points of the research methods. This project received ethical approval as a quality improvement project (University of Pittsburgh protocol #0001413).
