How do colorectal cancer patients rate their GP: a mixed methods study

Background New Zealand (NZ) has a high incidence of colorectal cancer (CRC) and low rates of early diagnosis. With screening not yet nationwide, the majority of CRC is diagnosed through general practice. A good patient-general practitioner (GP) relationship can facilitate prompt diagnosis, but when there is a breakdown in this relationship, delays can occur. Delayed diagnosis of CRC in NZ receives a disproportionally high number of complaints directed against GPs, suggesting deficits in the patient-GP connection. We aimed to investigate patient-reported confidence and ratings of their GP following the diagnostic process. Methods This study is a mixed methods analysis of responses to a structured questionnaire and free text comments from patients newly diagnosed with CRC in the Midland region of NZ. A total of 195 patients responded to the structured questionnaire, and 113 patients provided additional free text comments. Descriptive statistics were used to describe the study population and chi square analysis determined the statistical significance of factors possibly linked to delay. Free text comments were analysed using a thematic framework. Results Most participants rated their GP as ‘Very good/Good’ at communication with patients about their health conditions and involving them in decisions about their care, and 6.7% of participants rated their overall level of confidence and trust in their GP as ‘Not at all’. Age, gender, ethnicity and a longer diagnostic interval were associated with lower confidence and trust. Free text comments were grouped in to three themes: 1. GP Interpersonal skills; (communication, listening, taking patient symptoms seriously), 2. Technical competence; (speed of referral, misdiagnoses, lack of physical examination), and 3. Organisation of general practice care; (appointment length, getting an appointment, continuity of care). Conclusions Māori, females, and younger participants were more likely to report low confidence and trust in their GP. Participants associate a poor diagnostic experience with deficits in the interpersonal and technical skills of their GP, and health system factors within general practice. Short appointment times, access to appointments and poor GP continuity are important components of how patients assess their experience and are particularly important to ensure equal access for Māori patients.


Background
Trust and confidence in general practitioners (GPs) is usually reported as high [1]. Factors associated with patient confidence in GPs include clear explanations of tests and treatments, involving patients in decisions about care and patient perceptions that their symptoms are being taken seriously. When trust breaks down and care is perceived to be sub-optimal, conflict can ensue. In New Zealand (NZ) any complaint about health practitioners can be referred to the Health and Disability Commissioner (HDC). A report for the HDC (2004HDC ( -2013 indicated that approximately 10% of complaints about GPs involved a perceived delay in diagnosis of cancer [2]. Colorectal cancer (CRC) -the second most common cancer in NZ [3] -was overrepresented, comprising 27% of these complaints. The nature of complaints highlighted in the report were a lack of clinical examinations, patient perceptions of inadequate follow-up of symptoms and poor GP communication.
Based on international comparisons, NZ has a low rate of early stage CRC [4], with fewer than 12% of patients diagnosed at stage I [5].
While not yet having a fully implemented national bowel cancer screening program is a contributing factor, CRC is also difficult to diagnose [6], with a complex diagnostic process for both patients and GPs. For patients, symptoms can be nonspecific and difficult to recognise as potentially serious in nature and in need of medical investigation [7][8][9]. For GPs, the difficulty lies in the frequency of bowel symptoms in primary care and interpretive complications. Non-cancer diagnoses are much more common, and can include conditions such as irritable bowel syndrome, inflammatory bowel disease (Crohn's and ulcerative colitis) and diverticular disease. There is accordingly, significant potential for misdiagnosis, especially in the presence of comorbidity [10,11] or existing gastro-intestinal issues that can confound the presence of CRC symptoms [7,12,13].
The patient-GP relationship is an integral aspect of the diagnostic process. A GPs interpersonal skills (e.g., listening, empathy, being non-judgemental) and technical competence (e.g., knowledge, performing physical examinations, proactively investigating, following up on referrals) can either facilitate or impede prompt diagnosis. Good GP communication helps patients feel connected to their GP and the care provided [14], but a lack of empathy, inattentive listening and not taking patients seriously can lead to negative patient-GP interactions [15], patient dissatisfaction [9] and complaints [16]. Technical competence is also an important consideration in the patient-GP relationship, but can be outweighed by interpersonal competence [17], highlighting the importance patients place on a GPs' personal style during interactions.
Given the prevalence of CRC complaints in primary care in NZ, it is important to investigate patient reported confidence in their GP following the diagnostic process. We therefore interviewed patients recently diagnosed with CRC using a structured questionnaire to investigate factors that lead to high (or low) patient ratings of trust and confidence in their GP and how these factors contribute to the overall diagnostic experience.

Participants
Participants were selected from the Midland region, which includes Waikato (population: 400,000+), Tairawhiti (population: 40,000+) and Lakes (population: 100,000+) District Health Boards (DHBs). Participants were recruited as part of a larger prospective study where data were collected via researcher-assisted interviews that administered a 52-item questionnaire based on the SYMPTOM questionnaire [12]. Questionnaire data was then analysed based on the Model of Pathways to Treatment (MPT) [18] framework. Initial recruitment occurred through referral from a CRC cancer nurse specialist (CNS) at each of these DHBs and participants were contacted to arrange a time and day for interview. Additional recruitment within the Waikato region took place via mail out of study information from patient lists obtained from Waikato DHB, use of a poster placed at Waikato hospital and in private consulting rooms, and a Bowel Cancer NZ social media page. No interviews took place until a signed consent form or written consent via email or text message was obtained.
Participants were selected for recruitment if they had been diagnosed and interviewed within 12 months of diagnosis (study period from 2016 to 2019) and had not participated in bowel screening (where screening has been implemented regionally (e.g., Lakes DHB)). Interviews took place from April 2018 to March 2020. During the interview, participants were invited to speak about their experience of being diagnosed with CRC, with a particular focus on patient-reported symptoms and the timeline from symptom onset to when a health care professional (usually a GP) was consulted. The results of this larger study are not reported here. The current study reports on Section 3 of the questionnaire, which asks about health service utilisation and the patient-GP experience using three key questions (see Tables 2, 3 and 4).
Responses to these key questions were collected using a 5 -point Likert rating scale ranging from 'Very good' to 'Very poor' or 'Yes definitely' to 'Not at all' . All three questions also included 'Doesn't apply' and 'Don't know' as possible response options. In addition, free text comments were recorded verbatim by the researcher at any point during the interview, but were specifically prompted in Section 3 due to these questions' particular relation to the patient-GP experience. Ethical approval for this study was granted by the New Zealand Health and Disability Ethics Committee (Ref: 17/NTB/156).

Delay intervals
The MPT [18] provided the theoretical framework for data analysis and defines four intervals from first symptom/bodily change to commencement of treatment (appraisal, help seeking, diagnostic, and pre-treatment). This study reports only on the total diagnostic interval (TDI), defined as the date of first symptom onset to date of diagnosis as guided by the Aarhus statement [19] and a previous study [12]. Diagnostic intervals were defined as > 120 days or < 120 days, based on Australian clinical guidelines [20].

Data analysis
Descriptive statistics were used to describe the study population and the characteristics of the patients who provided free text comments. Chi square analysis was used to determine any statistical significance. All tests for significance were two-tailed with p < 0.05 considered a statistically significant result. All data analyses were performed using SPSS version 25 (New York, US). Additional free text comments were compiled, and these responses were analysed by the primary author (TB). As described in a similar study [11], free-text comments in the current study were also considered as unstructured and unguided qualitative data. Using an inductive approach guided by the data, thematic analysis techniques [21] were applied whereby comments were coded manually using highlighters in an Excel spreadsheet and then analysed for themes. Table 1 shows the characteristics of the current cohort. The participants were mostly aged 70-79 (33.8%), non-Māori (84.6%), male (55.9%) and had been diagnosed through their GP (64.1%). Over half (53.3%) of participants had a TDI of more than > 120 days. Dates needed to calculate TDI were unknown for 12 participants. Seventy-three (37.4%) participants could not get an appointment with a GP or nurse within 24 h of calling their medical practice for the purpose of making an appointment (for any reason). The main reason why an appointment could not be made was a lack of available appointments (24.6%). From the current sample, 113 (57.9%) participants offered free text comments relevant to GP-related care (characteristics are shown in the right hand panel of Table 1). Tables 2, 3 and 4 show how participants responded to three key questions relevant to the patient-GP relationship. Eighteen participants (9.2%) rated their GPs communication as 'Neither good nor bad/Poor or Very poor' . The majority (79.5%) of participants (n = 155) rated their GP involving them in decisions about their care as 'Very good/Good' . When asked for an overall judgment of confidence and trust in their GP, 40 participants (20.5%) rated that level of confidence and trust as 'Yes, to some extent/Not at all' .

Results
Chi-square analysis showed that age (p = 0.004) and gender (p = 0.028) were significantly associated with the confidence and trust rating. Proportionally, more Māori participants gave a 'Yes, to some extent/Not at all' rating of overall confidence and trust in their

Free text comments
Three themes were identified from participant free text comments: GP Interpersonal skills, technical competence and organisation of general practice care. For some, the perception of a poor GP relationship prompted a change to a different GP or medical practice. This was the case for two participants, who felt particularly dismissed by their GPs. One described a stressful 8 month 'fight' to get her GP to listen and initiate a specialist referral and the other felt totally disconnected to her care:  However, some participants were more accepting of their GP's interpersonal style, which did not affect confidence levels. One participant gave an honest description of his GP's poor communication, yet still had total faith in his care: He is terrible at explaining things. I have a long standing relationship with him, and even though he has quirky weird ways, he has proven his level of care to my family multiple times -when the chips are down, you can't beat him (Male, age 76, stage 3, TDI<120 days)

Theme 2: technical competence
A GPs technical competence was also typically appraised by participants during appointments, and provided the second theme identified. Technical skills were often judged by the speed in which a referral was made, which for some patients, was connected to levels of confidence and trust: Low confidence and feelings of dismissal were also evident in those participants who recounted being misdiagnosed in the absence of a physical examination: However, some participants still had confidence and trust in their GP, despite experiencing a longer diagnostic interval -especially if their GP was actively engaged in investigating symptoms, or participants could acknowledge that their own medical history was a significant contributor to diagnostic difficulty: One said I was 'too young for cancer' but still referred me, and did bloods (Female, age 31, stage 3, TDI>120 days) I have a history of endometriosis, so felt their assess-ments were fair (Female, age 37, stage unknown, TDI>120 days)

Theme 3: organisation of general practice care
While clearly beyond the scope of a GPs interpersonal and technical competence, many participants commented on health system issues within general practice, suggesting that some participants do not view these as distinctly separate from the patient-GP relationshipand in fact include these factors when assessing feelings of confidence and trust in their GP. Timing of appointments was a common concern, with short appointment times resulting in participants feeling rushed and not being given enough time for their concerns to be properly heard: Continuity of care was another main concern. While busy practices might offer an appointment with another GP, participants often desired to see the same GP who they felt they trusted:

Discussion
We investigated perspectives of the patient-GP relationship in the context of bowel cancer detection and diagnosis by analysing the free text comments and GP ratings from recently diagnosed CRC patients. A diagnosis of cancer is a critical time for a patient, in which expectations of general practice are high. Over half of the current cohort experienced a TDI of more than 120 days. Almost 30% of participants who had a long diagnostic interval gave a low rating of confidence and trust in their GP, suggesting that for these participants, a longer TDI was one of the contributing factors to this rating. Poor interpersonal skills, such as a lack of communication or listening, and poor technical competence, including misdiagnoses and not being thoroughly examined are factors that also impact on the diagnostic experience for patients and their level of confidence and trust in the patient-GP relationship.
While it was encouraging to see many participants rating GP communication positively, several participants voiced dissatisfaction with their GPs interpersonal manner, with some participants feeling 'desperate' to get their GP to listen, being made to feel like a hypochondriac, or left 'fighting' to be taken seriously. These feelings certainly contributed to a poor overall rating of confidence and trust for some participants. Patients value having their symptoms taken seriously [1], and want to feel that their GP understands their symptoms from their perspective [22][23][24]. This is especially important for patients disclosing often embarrassing CRC symptoms, and for Māori patients in particular, where revealing symptoms to an (often) non-Māori practitioner may be particularly difficult [25] -especially in the light of current inequities, where Māori have a lower incidence [26] but worse CRC outcomes [27,28], and less access to chemotherapy [29] and colonoscopy [30]. Consistent with other research [11], younger participants in the current study reported a sense of not being taken seriously and significantly more females reported low confidence in their GP. Young patients [13,31] and females [8,13,[32][33][34] are more at risk for delayed diagnosis, and the experiences of some of the participants in this study emphasize the importance of GPs not being dismissive of symptoms in these groups.
Participants also expressed dissatisfaction with the technical competence of their GP, commenting on the speed in which specialist referrals were made, often perceiving that their GP 'took too long to diagnose' . Misdiagnoses, especially in patients who experienced a longer TDI, were commonly reported by participants, and are a significant barrier to both patients seeking further GP consultations and GPs reaching a diagnosis [35]. However, accurate diagnosis of CRC symptoms is difficult [8,9], and GPs must interpret symptoms in the light of a number of factors, including the presence of comorbid conditions which may disguise CRC symptoms and increase time to diagnose [10,36]. NZ GPs are also disadvantaged by less direct and slower access to colonoscopy than GPs in other countries [37], largely due to a public hospital system that is based on triage for degree of need, resulting in patients who are not likely to be seen or treated within 6 months being routinely referred back to GPs without being seen. Of concern, however, were the participant accounts of misdiagnoses in the absence of a physical examination. Low rates of physical examination prior to diagnosis have been previously reported [6,8,34,38,39], and were one of the primary sources of complaint against NZ GPs [2], so are clearly an ongoing issue in CRC diagnosis in NZ, and one that contributes to a patients overall perception of the quality of their relationship with their GP.
Organisation of general practice care, while not under direct control or responsibility of many GPs, was another prominent theme. Half of the participants who provided free text comments reported an inability to access their GP for an appointment within 24 h over the preceding 12 months. Participants also commented on appointment length, feeling that a '10 min slot' was not long enough to have their issues heard. NZs standard 15 min appointment time is a funding issue, and has been raised as a point of concern by both GPs and primary care nurses [40]. Patients value GPs taking time during appointments [41], and do not like feeling rushed [23,42], therefore taking the time within existing consult times to carefully listen may help mitigate short appointment times. Clearly this is a balancing act for GPs. Getting an appointment with a desired GP was also highly valued. Irrespective of TDI, participants expressed frustration at not being able to see the same GP, or being offered a different GP for each appointment. Poor relational continuity of care, where a lack of consistency provides patients with unpredictability and no coherence [43], is a source of patient unhappiness [9], increases time to diagnosis [13], and makes patients feel like they are being treated impersonally [42]. This is a particular issue for Māori patients, who value continuity of care [44] but do not get offered the same choice of GP appointments [45]. We suggest that further investment is needed in primary care, and that primary care organisations focus on improving continuity and patient-GP communication.
Few studies have investigated the patient-GP relationship following a CRC diagnosis from the patient's perspective. We used a mixed methods approach to allow participant voices to be heard. Free text comments provide valuable additional data and are one way to measure a wider range of topics that might not be fully captured with a structured questionnaire [11]. However, these are not representative and so cannot be generalised to the views of all participants. Furthermore, patients with a CRC diagnosis are not typical of all cancer patients, so may experience the diagnostic pathway through general practice differently. Data collected was patient-reported, so relied on subjective memory of events and accurate recall of diagnostic dates. While interviews aimed to be conducted as close to diagnosis date as possible (at least within 12 months of diagnosis), patient recall may not have been accurate. Finally, while patient gender could be reported, GP characteristics (including age, gender, time in practice, practice size etc.) were unknown and would be important factors for inclusion in future research.
We report that long diagnostic intervals for CRC are still occurring in primary care, and that patients associate a poor diagnostic experience with deficits in the interpersonal and technical skills of their GP, and health system factors within general practice. Many of the issues reported here have been previously raised by an HDC report (2004-2013) [2]. Increased funding into primary care might help address some of these ongoing issues. While the majority of participants in the current study had confidence and trust in their GP, the diagnostic experience was extremely negative for some participants, particularly young patients, Māori, and females. Access to general practice plays a pivotal role and is particularly important to ensure equity for Māori patients.