Patient Clinician Information Exchange(PCIE)
There is a deep tension in the existing patient-clinician communication literature. On the one hand there have been a large number of studies that describe the patterns of communication with richness and rigor. On the other hand the literature is weak when it comes to establishing that variations in patient-clinician communication—whether observed or changed by deliberate intervention—are associated with improved health outcomes. In a sense this is inevitable; such rich assessments of communication interactions can only be done on relatively small, often convenience, samples with a focus on short-term results. For this proposal, we have started with this weakness in the literature and asked the following question: what can we do to assess the long-term effects of patient-clinician communication on health outcomes on a large representative sample of cancer patients?
The first issue we faced was what aspects of the patient-clinician communication variable could be assessed. The measures we have used to date have asked patients to describe memories of their interactions with clinicians at the time of diagnosis and have focused on the major elements of those interactions, rather than the subtleties that have been the focus of much of the extant literature. Also, although patients may have seen a variety of clinicians during their first year post-diagnosis, our instrument pretests made it clear that it would be unreasonable to expect them to report separately on communication with each clinician. Thus, we have focused on developing global measures of patient-clinician communication—global in the sense of averaging across treating clinicians, and global in the sense of averaging across multiple interactions. In this proposed study we focus on patient-clinician information exchange. Our measure, also named “patient-clinician information exchange” (or PCIE), primarily addresses the construct of information exchange while incorporating some aspects of both decision making and uncertainty management. We will focus on PCIE and its effects in the post-diagnosis, treatment, and survivorship phases of the cancer care continuum.
Our objectives direct the study in two dimensions: examining outcomes of PCIE and examining determinants of PCIE. We begin here with considerations of relevant outcomes and then turn to determinants. We focus here on outcomes which are particularly likely to be outcomes of PCIE rather than of patient-clinician communication more broadly, although in some cases it is difficult to discern precisely which element of communication is operating. Each set of outcomes can be seen as the mechanism through which PCIE affects the next level of outcomes. Relevant proximal outcomes include the accurate transfer and acquisition of information and the cultivation of trust; important intermediate outcomes include patient satisfaction with the consultation and treatment choice, knowledge and understanding, and coping with the stress of diagnosis, and adherence to treatment and lifestyle recommendations; and longer term health outcomes include reduced morbidity, and improved well-being and quality of life. There are prior studies supporting an expectation that higher levels of PCIE will be associated with reduced anxiety and mood disturbance, improved communication with family members, and decreased frequency of malpractice claims and patients’ ability to sustain hope. Recent reviews make it clear that the literature is better for establishing effects of PCIE on some proximal and intermediate outcomes, but weaker for longer term outcomes. To quote Epstein and Street “… very little is known about the mechanisms by which communication affects health outcomes”.
The proposed mechanisms of effects of PCIE on various outcomes in the literature are reasonably straightforward. Patients with richer information interactions are likely to be more satisfied with their relationship with their treating clinicians, have a better sense of being informed and more self-efficacy in managing their disease and its uncertain prospects; satisfaction is assumed to lead to a greater willingness to follow recommendations. These in turn might be expected to lead to regular surveillance visits, timely undertaking of diagnostic tests, adherence to medications, and acceptance of recommended lifestyle changes. PCIE would then be expected to affect health outcomes: Quality of life would be improved both because of the regular surveillance actions undertaken which might reduce symptoms, and because the sense of informedness and self-efficacy would give patients the confidence to deal with new issues that arose. The literature then argues for longer term reduced disease-specific morbidity (e.g., fewer recurrences) because the high PCIE patients have undertaken more of the actions recommended, and because they are more active, timely, and effective in engaging with the medical system when it is needed. In our current study, we will examine evidence for effects on a substantial number but still a subset of these outcomes.
The proposed research will elucidate the pathways that lead to patient clinician communication and allow us to determine the facets of patient clinician communication that lead to particular medical outcomes. By extension, this research will allow us to determine what types of patient clinician communication lead to more desirable medical and psychological outcomes (e.g., more effective treatments) and what variables predict engagement in desirable patient clinician communication. Others may then wish to use this information to facilitate particular types of patient clinician communication.
Back to top
Prior Research: Information Seeking and Scanning Behavior (SSB)
The Patient-Clinician Information Exchange (PCIE) builds on our unique sample survey undertaken under CECCR I with Pennsylvania patients diagnosed with either of breast, colon or prostate cancer in 2005. In the fall of 2006, the Pennsylvania Cancer Registry (PCR) provided a list of all cancer patients diagnosed with breast, prostate or colorectal cancer in 2005. Since reporting of cancer diagnoses by medical facilities is legally required in Pennsylvania, the PCR list is believed to include more than 95% of such patients in the state.
The questionnaires used in the survey included detailed questions about information seeking and scanning behavior; what medical, interpersonal and mediated sources were used with regard to treatment, other topics of concern at time of diagnosis and topics related to quality of life. In addition the questionnaires collected a wide range of background information, as well as possible outcomes of information seeking scanning, including psychological concerns, following recommendations, quality of life experiences and lifestyle behaviors.
Preliminary analyses of these data provide strong support for the relevance of active seeking by patients of information from their physicians to some outcomes.
Active information exchange is associated with positive regard for the perceived importance of the physician as an information source. However; the potential effects extend beyond physician assessment alone. In every aspect of their assessment of their self-efficacy with dealing with their disease and treatment satisfaction, higher information exchange was associated with positive responses. Those who actively exchanged information with their physicians believed themselves to be better informed about their disease, claimed that they were heavily involved in decisions around their treatment, were more likely to believe they knew what they faced in the future, had confidence in their ability to deal with future problems, and, beyond that, were more satisfied with their treatment decisions.
Information exchange with physicians is also linked to reports of engaging in medically-relevant behavior. Those who reported actively engaging in information exchange with their physicians were more likely to recall getting recommendations from their physicians and, if they received such recommendations, also were more likely to ‘always’ follow recommendations.
There was evidence that the patients who reported more active patient-physician information exchange also reported greater adherence with physician recommendations for tests, treatments and lifestyle changes. We wondered whether the reported compliance with lifestyle changes would show up also in actual lifestyle behaviors. The questionnaire asked about fruit and vegetable consumption, exercise frequency and dieting to control weight. Evidence from these preliminary analyses that respondents who reported more information exchange were more likely to be engaging in recommended lifestyle changes was mixed. Exercise behavior was not related to patient reports of active information seeking; however, both fruit and vegetable consumption and self-reports of past 30 day dieting to lose weight were related.
The specific aims of the SSB project in CECCR I focused on the role of public information sources in decisions; these complementary analyses focusing on the role of information contact with physicians were not part of the specific aims of the CECCR I project. However, these preliminary results offered some striking evidence for the potential importance of information exchange with physicians. There is clearly a great deal more to be learned from the existing data (including the follow-up round of data). Adding a third round of data collection in PCIE will enable us to observe longer case histories and examine continuing effects as people live with cancer over three to four years. Integrating and matching this self-report information with information from the Medicare claims files, which provides some parallel information and some additional information from the perspective of the reporting physicians, will substantially enhance what can be learned with these questionnaire data. These analyses and the additional data gathering will allow us to move from reporting intriguing cross-sectional associations to much more confident longitudinal-based assessments of influence.
Back to top
CECCR II: Outcomes of Patient Clinician Engagement
The overall goal of this study is to determine the relationship between measures of patient physician information exchange collected in the baseline Cancer Patient Survey and cognitive, behavioral and treatment outcomes assessed in the baseline survey as well as through subsequent rounds of the Cancer Patient Survey and links to Medicare claims files. In the end, the argument for PCIE is an argument that patients will fare better – face less morbidity or better adapt to and cope with their disease-- if they and their physicians engage in fuller communication.
We propose to examine the potential outcomes of patient-clinician information exchange, specifically whether patient-clinician information exchange is associated with and predictive of important health management decisions, behaviors and outcomes, including treatment choices, adherence with physician recommendations for tests and medications, lifestyle behavior changes, maintenance visits to physicians, and treatment associated morbidity.
The use of Medicare data provides several important advantages over the exclusively questionnaire-based analyses used in SSB. First, they provide the opportunity to obtain information about medical care and outcomes for a longer period of time than that encompassed in the survey data collection - from before diagnosis, during treatment, and during extended periods of follow-up. Second, they provide measures of potentially important and informative variables that are not available from the questionnaires. Third, they provide what some will see as ‘harder’ outcomes than patient self-reports on some parallel outcome variables– allowing us to see whether patient perceptions and histories can be validated against an external standard. Patients may not accurately recall behavior that took place a year or more before they filled out a questionnaire; they may not have learned to fully distinguish among the variety of specialists from whom they receive care and among the tests that they received; the particular phrasing of a question might not have drawn out the relevant experience. Demonstrating that effects observed from the questionnaire data are replicated when parallel or richer Medicare data are used will provide independent evidence of the validity of findings from the questionnaires.
Our two main objectives for Outcomes of Patient Clinician Engagement:
Back to top
To test whether patient-clinician information exchange during the initial cancer treatment phase is associated with subsequent differences in cognitive, behavioral and treatment outcomes in the first three years after treatment. Outcomes will be assessed through our established three rounds of cancer patient surveys from The Seeking and Scanning Behavior Project and linkage of these data to Medicare files. Our hypotheses are:
Patient-clinician information exchange is positively associated with contemporaneous and predictive of subsequent (i) patient confidence in their physician as a source of information, and patient (ii) self-efficacy in managing their disease, (iii) self-reported engagement in shared decision-making and (iv) inversely associated with concerns about the future (“cancer worry”).
Patient-clinician information exchange is positively associated with contemporaneous and predictive of subsequent healthy lifestyle behaviors and use of recommended preventive care including cancer surveillance testing.
Patient-clinician information exchange is associated with more aggressive treatment (including use of adjuvant chemotherapy in breast and colorectal cancer and use of prostatectomy in prostate cancer), lower rates of emergency room visits and hospitalizations for treatment complications, and higher levels of overall treatment satisfaction.
To test whether the effects of patient seeking and scanning of information from non-clinician sources, both personal and mediated, are larger when patient-clinician information exchange is limited:
The effects of patient seeking and scanning of information from personal (other cancer patient, friends, family) and mediated (broadcast, Internet, print) non-clinician sources on these outcomes will be greater when patients are engaged in a lower level of physician-patient information exchange.
CECCR II: Determinants of Patient Clinician Information Engagement
The overall goal of study two is to combine the available data from the three round study (baseline and two follow-ups) with Medicare administrative claims data, to explore what accounts for the differences among those who are relatively high versus low on PCIE. The study will take two approaches. First it will examine the issue broadly – using all of the potential confounders, from the PCR archive and selected measures from the Medicare claims files data to examine their relative influences on PCIE and on immediate outcomes of PCIE. Given preliminary data, where these relationships have been found, we expect that younger more than older, better educated more than less well-educated, and breast and prostate cancer patients more than colon cancer patients will report higher PCIE. We will also investigate the relationships of other patient demographic and clinical characteristics. Second the analyses will focus on the hypothesis that the ability of patients to interact successfully with specialist physicians, adhere with clinician recommendations and engage in healthy behaviors are enhanced if they have continuing contact with a primary physician or medical oncologist. Cancer treatment requires that patients obtain care from a range of specialist physicians. If, nonetheless, the patient has a physician with whom he/she maintains contact (whether a primary physician or a medical oncologist), it is hypothesized that the patient will be able to engage in PCIE more thoroughly and assess physicians and rate their own knowledge and self-efficacy to manage their disease more positively. There is no measure in the baseline instrument that will permit differentiation as to what sorts of clinicians a patient had seen for treatment (the generic term “treating doctors” is used to capture this set of physicians) or the frequency of such interactions. The Medicare data archive will permit assessment of patient physician utilization.
Assuming that patient-clinician information exchange is demonstrated to influence outcomes for cancer patients in a positive direction, policy makers will be faced with the question of how best to influence patient-clinician information exchange. Interventions that encourage active engagement by patients and/or target individual physicians to improve communication skills may be effective for some patient-clinician pairs - but may be cumbersome and inefficient to do on a large scale across a population of patients. In contrast, the identification of service patterns that predict high levels of PCIE may provide a relatively straightforward and efficient method of improving patient clinician information exchange. If certain patterns of physician visits with a primary care provider and different oncology specialists are correlated with higher levels of PCIE, policy makers can consider implementing policies to encourage or facilitate these patterns. If other patterns are particularly correlated with lower levels of PCIE, policies may be needed to discourage or create disincentives for these patterns.
Our main objective is to identify determinants of higher levels of patient-clinician information exchange. Potential determinants will be assessed through the cancer patient surveys and Medicare claims files. Our hypotheses:
Back to top
Patients who have continuing visits to a single primary care clinician during their initial cancer treatment phase report higher levels of patient-clinician information exchange and greater confidence in their physicians as a source of information.
Patients who have continuing visits to a single medical oncologist during their initial treatment phase report higher levels of patient-clinician information exchange and greater confidence in their physicians as a source of information.
Patient-clinician information exchange will vary with patient demographic characteristics (younger more than older, better educated more than less well educated) and with patient clinical characteristics (breast or prostate more than colon).