Previous evidence regarding care transitions between secondary or tertiary and primary health care has shown that receipt, timeliness, and content of communication is often suboptimal [20–24]. This study investigated early shared care communication processes between CR and PCPs. Similar to Riley et al.'s study which showed only 42% transmission of CR discharge transition records to PCPs , results herein revealed that less than half of PCPs received an intake transition record for their patients. There were significant differences in transmission rate by CR program, suggesting lack of standardized processes. However, given that almost all PCPs want to receive an intake record so they have a comprehensive summary of their patient's health which they can track over time, clearly standardization is warranted. Moreover, most PCPs were not aware their patients had enrolled in a CR program until they were notified by the patient. Overall, PCPs prefer to receive a concise (less than 1.5 page-long) faxed intake transition record that includes clinical status, exercise test results, and risk factor status in relation to evidence-based targets.
CR Intake Transition Record Transmission
PCPs were quite satisfied with the timeliness with which they received the transition records. However, less than half of PCPs received the intake transition records. Moreover, there was wide variation in PCP receipt of CR intake transition records based on CR site, ranging from 0-100%. The lack of PCP receipt of CR intake transition records can be partially explained by one CR site only sending an intake transition record if there was no follow-up from an in-house specialist, and another CR site not sending CR intake transition records at all. As well, only approximately three-quarters of intake transition records that CR programs confirmed were sent to PCPs, were received. This may have been due to the practice at some CR programs to send the transition records only to the specialist who refers the patient, rather than identifying all relevant providers .
This reality contrasts sharply with the results of this study and previous work by our group  which suggest that almost all PCPs want to receive CR transition records. Over three-quarters of PCPs report they use the information within the record to manage their patients' risk. In particular, PCPs used the information for managing their patient's medications, followed by exercise habits and weight. The PCPs use the information as an indicator of patient enrollment, to track change in clinical and risk factor status, and to ensure they have a comprehensive summary of a patient's health. The latter likely provides impetus for the initial referral, as CR programs offer inter-professional care which cannot be replicated by a PCP alone. Clearly, professional and community standards need to be developed, and processes should be established to promote effective transition communication [8, 15].
Over three-quarters of PCPs desired transition records via fax, and only one-quarter desired them via electronic means. None of the CR programs transmitted via fax, and this would likely increase timeliness of PCP receipt. Format incompatibility of information sent by CR and stored in PCP offices was identified as a barrier to optimal transitional care. At this time, many PCPs are in the process of adopting EHRs and/or prefer hard copies of intake transition records. The cardiac healthcare community needs to develop standard transmission protocols to optimize transitional inter-provider communication.
CR Intake Transition Record Content
Interviews with PCPs enabled feedback regarding their needs for transitional care. PCPs were highly satisfied with both previous CR transition records received, as well as with the quality of the intake records received for patients in the study. There was, however, a noticeable range in the degree of satisfaction of intake reports by CR site. A number of possible reasons could explain this variation, which may include the size and type of CR site, the site location, or even differences in the presentation of the intake reports. To pinpoint the reasons for this variation requires a more robust study design (i.e. a cluster randomized trial). In addition, PCPs were highly satisfied with the organization of the intake reports, and perceived they could easily understand all the content as presented. Overall, the transition records met PCP needs, made PCPs more willing to refer future patients, and were often used by PCPs for reference.
The PCPs expressed a desire for concise (less than 1.5 pages) intake transition records. PCPs were highly satisfied with the length of transition records received, and the time required to read them. Data elements that PCPs consider as most pertinent to include in intake transition records were the clinical status of the patient, exercise stress test results, patient care plan, and the medication list. Similar findings were reported in studies examining discharge transition records [15, 25, 26], where the clinical diagnosis and list of medications taken were considered more important details when compared to other data elements. Elements that were considered least important to PCPs were non-modifiable risk factors and subjective reports of the patient's level of motivation or emotions. However, the recent Transitions of Care Consensus Policy Statement, recommended that a section of the transition record be devoted to communicating "a patient's preferences, priorities, goals and values" . Whether this information is perceived as unimportant by PCPs because they rely on their own interactions with patients or patient reports as a source of this information deserves future investigation.
Other Patient-level, Clinical and Policy Implications
Over 80% of PCPs reported that CR intake transition records should also be provided to the patients themselves, in lay format. Indeed the Transitions of Care Consensus Policy  advocates that patients, families and caregivers should have secure and private access to their information ("nothing about me, without me"). Patients should receive, understand and be encouraged to participate in the development of the transition record. They also specify that this process should take into consideration patients' health literacy and be culturally-sensitive. Providing patients with a summary of their baseline health status would empower patients to set their own goals both inside and outside the program, and may ultimately improve self-management. Empowering the patient with the tools to self-manage promotes active patient engagement in their health care, and along with support from PCPs, has been found to improve management of chronic health conditions and reduce health care costs .
Caution is warranted when interpreting results, chiefly due to response bias. First, female PCPs and those graduated from medical school more recently were significantly more likely to participate in the study than older, male PCPs. Second, given the low PCP response rate, the representativeness of the sample cannot be firmly established. However, we compared the sex of PCPs from our sample to the Canadian Medical Association master files (available online at http://www.cma.ca/index.cfm/ci_id/16959/la_id/1.htm#1) and found no significant differences in the percent of females between the study sample and national data (44.4% vs 38.5%; p = 0.17) [data not shown]. Overall, the relative homogeneity of physician samples has been established , lending credence to the generalizability of the findings. However, ultimately we cannot rule out the possibility that respondents had more favourable views towards CR.
Second, CR programs and PCPs were located in Ontario where some CR services are partially reimbursed through government, which may limit generalizability to other reimbursement systems (or lack thereof). However, the CR sites in this study were chosen to reflect a diversity of patient volumes, number of CR staff, length of program, degree of rurality, and whether the program was hospital or community-based to mitigate against this threat to generalizability. Finally, while each of the CR programs participating in this study provide services in accordance with Canadian guidelines , we cannot rule out diversity among programs which may influence providers' perspectives.