Measuring cancer care coordination: development and validation of a questionnaire for patients

Effective coordination of care between different clinicians, services and health sectors throughout the patient journey is fundamental to the provision of high-quality care [1–3]. In health systems where care is well coordinated, patients will experience effective flow of information between clinicians throughout the course of their illness, with streamlined service provision in response to their physical, emotional and social needs [4]. Not only is good care co-ordination essential to optimize patients' experience, but it has also been shown to reduce future need for supportive care and to improve psychosocial outcomes [5].

People with cancer are particularly at risk of receiving poorly organized and fragmented care due to the complex nature of the disease and its management, which often involves multidisciplinary care from a large team of medical, nursing and allied health practitioners in both hospital and community settings over extended periods of time. As a result, many national strategic cancer plans have identified the improvement of cancer care coordination as a priority for service improvement [1, 4, 6, 7].

Efforts to improve cancer care coordination to date have been hindered by a dearth of accurate and reliable measures by which progress can be monitored. This partly stems from the lack of an agreed theoretical framework or definition of the term 'care coordination' to underpin the development of measures. For example, a recent literature review prepared for the Agency for Healthcare Research and Quality (AHRQ) identified more than 40 different definitions for 'care coordination'. However, the authors identified a number of common elements to inform the following working definition:

'Care coordination is the deliberate organization of patient care activities between two or more participants (including the patient) involved in a patient's care to facilitate the appropriate delivery of health care services. Organizing care involves the marshaling of personnel and other resources needed to carry out all required patient care activities, and is often managed by the exchange of information among participants responsible for different aspects of care [8].

This definition provides a starting point to identify the specific aspects of the care experience that should be addressed in any measurement tool.

There are a number of sources of data that could be used to assess aspects of cancer care coordination, including administrative health datasets, audits of individual patient records or measures based on the experience of patients or clinicians. Patients are ideally placed to rate the adequacy of cancer care coordination, however, as they are likely to be the only individual present at every encounter with health services. Furthermore, the move towards more patient-centred care in which services are organized around the needs and preferences of individual patients, emphasizes the primacy of measures based on patients' own experience. Therefore we conducted this study to develop a questionnaire for patients to assess their experience of cancer care coordination in the treatment phase of the cancer journey, to define the underlying factor structure of the questionnaire and to conduct initial validation by assessing construct validity, internal consistency and test-retest reliability.

Despite increasing recognition of inadequate care coordination as a common problem experienced by patients, to date there have been few measures by which improvement or deterioration in this crucial aspect of cancer care could be measured. The aim of this study was to develop a valid and reliable self-administered questionnaire for patients to measure the adequacy of cancer care coordination for those in the treatment phase of the cancer journey. The resulting questionnaire demonstrated robust psychometric properties and consistent subscales, suggesting that this instrument could provide a useful tool to measure cancer care coordination in future patient surveys and intervention studies.

The process of developing a new questionnaire is lengthy, requiring a number of iterative steps to further refine the wording and items to provide an instrument that is both acceptable and easily-understood by the target audience as well as providing a comprehensive, accurate and reliable measure of the phenomenon of interest. Furthermore, there is always a tension between the comprehensiveness of the instrument and the burden that a lengthy instrument will place on respondents. Brief instruments may achieve higher response rates, but may also limit the breadth or depth of information that can be collected. The approach to instrument development in this study was to only include items that had sound psychometric properties. Unreliable items that elicit inconsistent responses from an individual are of no value, as are items that are frequently missed out, perhaps due to lack of clarity in the wording or lack of relevance to a significant number of individuals. Furthermore, items that elicit highly skewed responses, with almost everyone giving a similar response, are of limited value for measurement. On the basis of these considerations, the draft questionnaire was reduced to 20 items that demonstrated good internal consistency and addressed two important components of cancer care coordination, namely the issues of communication and navigation of the health care system.

The Communication factor was the strongest, accounting for nearly 75% of the variance and demonstrating internal consistency in the desired range of 0.7-0.9 [16]. Comprising fewer items, the Navigation subscale necessarily had a lower value of Cronbach's alpha as this is partly dependent on the number of items in the scale [16]. Although the response distributions for the total score and Communication subscale were approximately Normal with no evidence of a ceiling effect, this was not the case for the Navigation subscale. There was a marked ceiling effect for this subscale, suggesting it may have limited usefulness as a stand-alone measure. Further development of this subscale, through inclusion and testing of additional items is warranted.

'Care coordination' and 'continuity of care' are related but distinct concepts [21–23]. While 'care coordination' broadly addresses process issues relevant to streamlined and appropriate navigation of the health care system, 'continuity of care' focuses more on consistency of information and clinical management between providers and over time, and on continuity within relationships [21–23]. As a result, measures of continuity of care have often focused on the issue of whether a patient saw the same doctor at each follow up visit [10, 24, 25]. Although seeing the same doctor is desirable in certain circumstances, for example within a specific clinic, this aspect of care is less relevant for a broad assessment of the coordination of cancer care which is often multidisciplinary in nature, involving consultations with a number of different health professionals where good communication and exchange of information is paramount. Other existing questionnaires have focused on assessment of patients' experience of hospital discharge, in recognition that the risk of poor care coordination is particularly high at times of transition in care [11, 26].

In the non-oncology setting, McGuiness and Sibthorpe took a broad approach to measuring health care coordination, developing an instrument for older patients with chronic, complex medical conditions in the primary care setting [12]. Others have included a single or small number of relevant items pertinent to care coordination within questionnaires to assess perceptions of the quality of care or satisfaction with cancer treatment generally, however this approach limits the depth of information gathered specifically about care coordination [27, 28] In contrast, our instrument was designed to provide a more comprehensive assessment of cancer care coordination based on the issues identified by patients and clinicians in our previous qualitative research.

Overall, rates of item completion were high, suggesting that the questions were clear and acceptable to patients. Of note, the highest rates of missing data were for items asking about family and carer issues. The reason for this warrants further investigation as it could be that some patients do not have a carer, or do not identify their family or friends as 'carers', or are unable to answer questions relating to the experience of their carers. Assessment of the experience of cancer care coordination from the perspective of carers warrants further research.

A number of limitations to this study are acknowledged. Although our sampling strategy aimed to include a broad range of patients in Sample 1, there was a preponderance of those with colorectal cancer and people from metropolitan centres. Furthermore, people with limited English skills were excluded from both samples and so the questionnaire may not be applicable to those from culturally and linguistically diverse communities. The methods used in this study were to reduce the number of questionnaire items to those with good psychometric properties and to provide a brief questionnaire. It is possible that the resulting instrument omits important aspects of cancer care coordination and the development of additional items and subscales could improve the content validity of the instrument. Furthermore, the responsiveness of this instrument to change needs to be tested in future studies.

In conclusion, the questionnaire developed in this study has been shown to be a psychometrically robust patient-report measure of cancer care coordination. Further studies will help establish the usefulness of this measure in future needs assessment surveys and intervention studies.