A survey of cancer patients’ unmet information and coordination needs in handovers – a cross-sectional study

Study population

The study population comprised all patients attending the two outpatient clinics described above in the period 2 January – 1 March 2012. Patients were eligible if they were 18 years or older and if they were more than three months post-diagnosis. Patients were excluded if a research nurse (RN) assessed that patients were unable to understand and complete a Danish questionnaire or were judged to be cognitive impaired. Eligible patients who came for a consultation were informed about the study by an RN and given the questionnaire. Participants returned the completed questionnaire in a prepaid return envelope together with a signed letter of consent. Information on sex, age, years since diagnosis, cancer type and treatment were retrieved by RNs from medical records.

Questionnaire

The self-administered questionnaire comprised 50 items within the domains of information and coordination and seven items related to sociodemographic characteristics. The items on information and coordination focused on various handover situations (handovers between primary and secondary sector, between internal hospital departments and between hospitals). Furthermore, the questionnaire included items of clinical relevance to outpatient clinics, and patients could add supplementary comments.

Respondents indicated their level of needs on a four-point Likert scale with the following response categories: “Strongly agree”, “Agree”, “Disagree”, and “Strongly disagree”. The responses were coded with the values one (“Strongly agree”) to four (“Strongly disagree”). Furthermore, a “Don’t know/Not relevant” response option was given. In order to handle missing values, a patient score was calculated as the mean of the responses corresponding to the eight scales of needs. In the analyses each patient’s score was dichotomised, using a score of two as cut-point. A mean below or equal to two was equivalent to “Strongly agree” or “Agree”. Likewise, a mean above two was equivalent to “Disagree” or “Strongly disagree”. A score less than two was categorised as having “no unmet needs” and a value equal to or greater than two was categorised as “having unmet needs”.

The items regarding information formed three handover specific subscales; hospital discharge (eight items), handovers between internal hospital departments (ten items), and handovers between hospitals (nine items). The 27 item together formed an overall information scale.

Development and piloting of patient questionnaire

The questionnaire was developed through different approaches. Classification of needs in the two domains was made based on a comprehensive literature search of needs of cancer patients. Relevant items from various existing instruments for measuring needs of cancer patients [6, 13, 18, 23–27], were adjusted to apply in the context of this study. Researchers in the field of cancer care and external reviewers assessed the contents, clarity of wording, and comprehensiveness of the items [28]. Three pilot tests, including a total of 25 cancer patients, were conducted. Piloting included time used to fill in the questionnaire and interviews with the patients about comprehension of the items [29]. Based on the three pilot tests the questionnaire was revised. A final pilot test, including hand-outs of the questionnaire to 63 cancer patients, tested the hand-out procedure and the willingness to participate (response rate of 67%). Patients from the pilot tests were not included in the final study sample.

Statistical analysis

Descriptive statistics were calculated, comparing participants with non-participants. This study considers a total of eight different outcome variables concerning needs in handovers. Four outcomes relate to information needs (three subscales as well as overall information need), and four outcomes relate to coordination needs (three subscales as well as overall coordination need). Univariate logistic regression was used to analyse the crude associations between each outcome and each of the independent variables comorbidity, treatment type, time since diagnosis, gender, age, and education, respectively. Education was dichotomised as lower education (less than three years of continuing education), and higher education (three years or more of continuing education).

Due to the limited number of participants it was only possible to adjust for a small number of potential confounders. It was decided to adjust for age, gender and education as the literature search identified these as important potential confounders. Cancer type was not included as a confounder, since some of the cancer-diagnoses in the study are gender-specific. In order to adjust for the three potential confounders, we conducted separate multiple logistic regressions analysing the associations between each of the outcomes and each of the independent variables, adjusting for age, gender and education. Gender and age are known to be associated with patient needs, as women and younger patients are generally more likely to report unmet needs [13, 19–21]. The patients’ levels of education has shown to influence rehabilitation and information needs [19–21].

A p-value below 0.05 was considered statistically significant. Internal reliability was assessed using Cronbach’s α and Pearson correlation. EpiData (The EpiData Association, Odense, Denmark) was used for entering data and all statistical analyses were performed using Stata 12 (StataCorp LP, College Station, Texas, USA).

Ethics

The study was approved by the Danish Data Protection Agency and the senior consultants at the Oncology Department and at the Surgery Department at Hospital Southwest Jutland, Esbjerg. According to the Danish Research Ethics Committee System we did not need the committee’s approval [30, 31].

Patient characteristics

Prevalences and associations

Hospital discharge was the handover that generated the highest level of unmet information needs (18%) compared to handovers between internal hospital departments (13%) and between hospitals (12%) (cf. Table 4). The highest level of unmet coordination needs was reported in handovers between hospitals (19%) and between hospital and GP (18%). Only few patients (8%) indicated unmet coordination needs in handovers between hospital departments (cf. Table 5).

Patients who underwent surgery or received chemotherapy had higher odds of reporting information and coordination needs than patients who did not undergo surgery and did not receive chemotherapy.

No general tendencies were observed in relation to gender, comorbidity, time since diagnosis, radiation therapy, and hormone treatment.

Main findings

Approximately every sixth patient reported unmet coordination needs and 18% reported unmet information needs. Hospital discharge was the handover with the highest level of unmet information needs. Both in handovers between hospital and GP and handovers between hospitals unmet coordination needs were relatively frequent. Age and education were statistically significantly associated with reporting unmet needs in both of the overall scales and in several of the subscales. Thus, patients younger than 60 years and patients with a higher education were more likely to express unmet needs than patients older than 60 years and patients without a higher education, respectively.

Strengths and weaknesses

A strength of the study is that the final questionnaire was developed based on previously used instruments for measuring needs of cancer patients and it had been thoroughly pilot tested. Furthermore, Cronbach’s α showed high internal reliability of the scales for each domain. However, some of the Cronbach’s αs were very high (α > 0.95) suggesting that there might be redundant items in these scales. Cronbach’s αs do not provide a thorough assessment of validity or general reliability thus a lack of rigorous psychometric evaluation of the survey can be considered a weakness.

Registering eligible patients and introducing them to the survey plus collecting journal information meant extra work for the already busy RNs. Thus an outpatient setting might not be optimal for a survey of this type. In order not to let the surroundings (i.e. the busy outpatient clinic) influence responses patients were recommended by the RNs to fill in the questionnaire at home. This however did not allow for sending reminders to potential participants which might have improved the response rate.

It is reasonable to interpret the indication of disagreement as an unmet need as the questionnaire use the wording “sufficiently”. However the questions used in the study do not directly address whether the patients actually wanted help with aspects of information and coordination.

A comparison of participants and non-participants revealed only minor differences. Breast cancer was the most prevalent diagnosis both among participants (71%) and non-participants (69%) exceeding the national breast cancer prevalence of 23% among Danish cancer patients [32]. This is partly explained by the specialisation of Danish hospital departments. Thus, treatment and follow-up of patients suffering from other cancer types in the area of Esbjerg often take place at other hospitals. Among participants patients three years post diagnosis were underrepresented compared to non-participants. This might be due to patients not wishing to revisit earlier phases of their cancer care trajectory or believing that their experiences were outdated and thus not relevant for the study.

Our study included a combination of questionnaire data and data obtained from medical records. It would have been interesting to also include data on e.g. depression, because this characteristic may be associated with unmet needs [33].

The statistical analyses adjusted for some important confounders. However, for some of the analyses the number of participants was limited resulting in reduced statistical precision reflected in the wide CIs.

As this study included patients in different phases of their cancer care trajectory, and as there are variations in the experiences of cancer patients, the recollection of experienced needs may vary. Generalisation of the results to other patients groups should be made with caution, because research has shown that patterns of needs depend on the phases in the care trajectory [11, 21, 22]. However, it is reasonable to assume that the tendencies indicated by the present study can be guiding for unmet information and coordination needs in similar contexts.

Comparison with other studies

Our study found that the highest level of unmet information needs was reported in relation to hospital discharge. This is in agreement with a Danish study which indicated that patients discharged experienced a lack of information and that the information given at the discharge was inconsistent between patients. Additionally, patients requested personalised information on possible side effects and wished to be treated by only a few different healthcare providers [34].

Knowledge of what patients find important or beneficial in care coordination is sparse [35]. Studies relating to coordination of care are often focused on cooperation which is concerned with the structural organisation, or on patient preferences for care and involvement in the decision-making process. Despite increased focus on care coordination, a recent Danish study of outpatients’ experiences (n = 96,860) indicated that there is still substantial room for improvement in relation to handovers [36]. The study found that 15% of patients experienced that their GP was not well-informed about the treatment of the patient. Approximately 7% experienced poor coordination between hospital departments and between hospital and municipality. Although these findings are not cancer-specific and are concerned with cooperation, rather than coordination and information, they support our results. Likewise, Grønvold et al. (2006) conducted a comprehensive Danish study (n = 1,518) of the needs and experiences of cancer patients. They found that a large proportion of cancer patients did not experience optimal cooperation between hospitals. Some 20% of the patients did not experience satisfactory cooperation between hospitals and 20% did not experience optimal cooperation between hospital and GP. Furthermore, patients did not find information to be timely, and frequently requested additional information. In accordance with our study, patients younger than 60 years old and patients with a higher education reported a higher level of unmet needs [18]. The age variation in needs might be explained by today’s younger patients being more explicit about their specific needs [19]. Furthermore, younger patients are more critical towards authorities, demanding dialogue, respect and good service [37]. As cancer prevalence is higher among the elderly, younger patients are less likely to have age-related peers going through a similar situation. This in itself could be isolating and reduces the opportunity of younger patients to gain insight and information from peers in a similar situation. Thus, it is possible that younger patients have different information and coordination needs than older patients. Earlier studies have shown that patients with a higher education have different information-seeking patterns than those without a higher education. Patients with high education often find alternative information sources [19] and are more specific when requesting information [21].

We did not observe associations between gender and the experience of unmet needs. Other studies have demonstrated strong gender differences in the experience of needs, where women are more likely to express unmet needs [18, 19, 38]. Moreover, comorbidity has been shown to be associated with a high level of rehabilitation needs [20]. Surprisingly, in our study patients without comorbidity had a tendency to report unmet needs more frequently. A reason for this could be that patients with comorbidities may have had previous experiences with the health care system and might thus know how to handle needs. Furthermore, information on comorbidities in this study was obtained from patients’ self-report and could be underreported [39].