Doctor, What’s Wrong with Me? Factors that Delay the Diagnosis of Colorectal Cancer

Abstract

Objective

To examine the influence of patient and physician communication factors on diagnostic delay (DD).

Methods

242 patients diagnosed with colorectal cancer (CRC) in the past 6 months who experienced symptoms prior to diagnosis were administered a 2-hour semi-structured qualitative interview to assess communication with health care provider and ease of access to care, among other factors. Patient-provided information was verified via review of medical records.

Results

The factors associated with DD > 2 months included lower income (OR=0.56, p=0.03), having regular physician prior to receiving a cancer diagnosis (OR=2.52, p=0.03), having a physician who used temporizing communication strategies during the consultation (OR=2.41, p=0.02), receiving an initial alternate diagnosis (OR=3.36, p=0.02), experiencing referral delay (OR=3.61, p=<0.001), and experiencing follow-up delay of any kind (OR=3.32, p=0.01).

Conclusion

Excellent communication skills that appropriately probe for relevant social and economic patient information, assist patients in distinguishing and elaborating on symptoms, and provide clear rationale and instructions for future steps, will speed along the diagnosis process and could be the difference between early and late stage CRC.

Practice Implications

Increased understanding of physician communication and practice styles that contribute to DD could have a positive impact on decreasing the morbidity and mortality from this disease.

1. Introduction

Colorectal cancer (CRC) is the second leading cause of cancer death in the United States with an estimated 142,570 new diagnoses and 51,370 CRC deaths in 2010 [1]. Unlike other cancers, the benefits of early detection and screening are clear. When detected early, the 5-year survival rate is 91% [1]. Therefore, it is imperative that the factors contributing to delays and disparities in CRC detection and diagnosis be understood.

Diagnostic delay (DD) is an important factor to consider regarding cancer disparities. DD can occur during the period when the patient first presents with symptoms to the formal health care system to the time of CRC diagnosis. DD is reportedly significantly and negatively associated with cancer patient outcomes [2–5]. In general, substantial barriers to early detection and diagnosis include lower rates of screening by minorities [6–8] and access issues due to lack of health insurance [9–11]. While most studies have examined breast cancer or a mix of cancers, several have been specific to CRC. CRC DD in a French sample was not associated with geographic or demographic factors such as gender, marital status, or type of physician first consulted [12]. In Denmark, lack of knowledge about the significance of bowel symptoms was a main component of CRC patient delay [13]. A similar lack of knowledge was found in a survey of 1332 randomly sampled adults in Australia [14]. A German multifactorial study of CRC DD found that patient occupation, education, and type of health insurance play a significant role, as does the clinical experience of the physician first consulted [15]. A U.S. study of psychological factors related to delay (n=69) reported that two-thirds of CRC patients did not understand the meaning of their symptoms; anxiety and harm avoidance were moderately correlated with delay [16]. Finally, a study of Veterans Affairs CRC patients concluded that ‘how physicians asked’ for colonoscopy was important, which included downplaying the potential severity of the symptoms [17].

Accumulating evidence documents cultural factors, including communication, that may play a role in the quality of care received and outcome disparities. Communication is key to the diagnostic process; physician decisions are influenced by their understanding and interpretation of their patients’ presentation of symptoms. Suboptimal doctor-patient communication may, therefore, influence DD. Patient-physician communication can be impeded by patients’ abilities to describe symptoms or their attitudes towards physicians [18]. For example, factors that influence the perception and reporting of symptoms, such as a belief that certain bowel habit changes are normal as one ages or fatalism, may impede an earlier cancer diagnosis in racial and ethnic minorities. In addition, the gender of the communicators can affect interactions [19]. These factors may hamper communication between physicians and patients, reducing the effectiveness of their communications. Other factors, such as patients’ socioeconomic status or health literacy level, which have received less attention by researchers, may also play a role. In one study, patients who had fewer than six years of formal education, were older than 50, or had a poor prognosis, received less thorough information [20]. Low income patients and those with less education ask fewer questions, which can signal their doctors to provide less information [21]. Finally, physicians who do not use affective and partnership behaviors tend to discourage patient participation in the medical encounter and decision making [22–24]. Affective communication is the process through which people express feelings about things, themselves, and others by, for example, exploring emotions or making a personal connection through shared laughter. Partnership behaviors invite patients to participate in the consultation and serve to flatten status and power differentials between physician and patient. These behaviors or their absence may impact physician interpretation and response to symptom information.

The objective of the present study was to describe the communication and symptom appraisal strategies of CRC patients and determine the patient and physician characteristics, consultation communication factors, follow-up recommendations, and the referral delay factors associated with DD.

2. Methods

2.1 Participants and recruitment

Participants were patients of any age or gender diagnosed with Stage 1 to 4 CRC who noticed symptoms indicative of CRC prior to diagnosis. Those individuals whose diagnoses occurred through routine screening colonoscopy were excluded from participation. Patients were recruited from 5 academic and community oncology practices in the states of Virginia and Ohio.

This study was approved by all relevant Institutional Review Boards and all participants provided written informed consent. All potentially eligible patients who received a CRC diagnosis in the prior 6 months were identified through prospective chart reviews at participating practices (n=458). Patients were then screened by telephone to determine if they reported experiencing symptoms prior to diagnosis and met eligibility criteria (n=303). Those who were eligible and consented to the study (n=256) participated in a 2-hour semi-structured interview. All relevant medical records concerning the diagnostic process leading up to the final CRC diagnosis were reviewed. The results presented here are based on analysis of a sample of 242 patients with complete data (interview data and data abstracted from medical records review) who were interviewed an average of 4 months after diagnosis.

2.2 Measurement

2.2.1 Patient interviews

A semi-structured cognitive interview was conducted that addressed: (1) symptom recognition and appraisal, (2) influence of family and friends on healthcare seeking behaviors, (3) social support, (3) ease of access to care, (4) communication with health care provider, and (5) sociodemographic factors. Interviews ranged from 45–210 minutes and consisted of a structured set of open-ended questions with specific follow-up probes, as well as closed-ended questions.

Because these types of interactions between physicians and patients are unpredictable and mostly occur within the context of primary care settings, direct observation is not feasible on a large scale. An indirect measure of the relational aspects of the physician-patient encounter was therefore obtained by administering the Relational Communication Scale (RCS) [25]. The RCS uses self-report from either one or both communicators to measure the interpersonal characteristics of the communication event. This measure consists of 8 sub-scales: Immediacy/Affection, Similarity/Depth, Receptivity/Trust, Composure, Formality, Dominance, Equality, and Task Orientation. The RCS is commonly administered to a single communicator and has been found to have acceptable validity and reliability [25], even by researchers using different versions [26, 27].

The open-ended interview questions were designed to elicit the patient’s story of his/her symptom appraisal and the evolution of the diagnostic process. The interview was transcribed and a coding sheet was developed to quantify the patients’ responses to each open-ended question. The questions and codes that were developed for this study were based on Arthur Kleinman’s concept of Explanatory Models [28, 29] and the Communication Model of Shared Decision Making [30]. For instance, constructs such as whether or not the physician misunderstood the patient’s symptoms and whether or not the physician believed the symptoms to be serious were coded from the patient’s narrative. Communication data collected refer to the patient’s first encounter with their health care provider regarding their symptoms. The codes were generated from iteratively coding interviews until thematic saturation was reached. Trained coders applied the operational definitions of each code, according to the coding manual. Double coding was conducted on 20% of cases through a process of inter-coder consensus; any coding discrepancies were resolved at regular coding meetings. Patient self-report data about physician encounters was verified via review of medical records.

2.2.2 Diagnostic delay

DD was defined as the time from the first medical visit in which the patient reported symptoms indicative of CRC through the date of diagnosis. These dates were obtained initially from patients using a calendar review process and confirmed through review of medical records. Clinically significant DD was defined as delay >2 months. The literature [31–32] suggests that 2 months is an appropriate DD threshold for a patient presenting with vague or ambiguous symptoms to be followed-up with testing (e.g., blood work) and have had a colonoscopy scheduled or performed.

2.2.3 Statistical Analysis

Univariate analyses between key variables and a log-transformed continuous DD variable were conducted using t-tests or ANOVAs with categorical variables and non-parametric Spearman’s correlations with continuous variables. Alpha was set at 0.05 for these analyses. Logistic regression was performed using the dichotomous outcome variable DD. Predictors with a statistically significant relationship with DD in the univariate analyses were considered in individual regression models (results not reported). Significant factors were used for the final logistic regression model, in addition to sociodemographic factors. All predictor variables were entered simultaneously into the logistic regression model. Only the final model results are reported.

3. Results

3.1 Sociodemographic characteristics

Of the 242 patients interviewed, 52.1% were male and 43.0% were African American. The average age was 58 years, 53.3% were married, and 52.1% had more than a high school education. Almost half of the participants (45.5%) were employed, 59.1% earned less than $50,000/year, and 45% had private health insurance. The majority (69.0%) were recruited from Virginia. Most patients (64.9%) had Stage 2 or 3 CRC at diagnosis. This reflects the entry criteria for the study. The mean DD experienced was 4.8 months (SD=8.3), ranging from 0–55.7 months, and median DD was 1.5 months. 122 (55.0%) participants had DD >2 months. Table 1 displays the sample’s sociodemographic information.

Table 1.

Sociodemographic characteristics of the sample (n = 242)

Characteristic	n (%)
Gender
Male	126 (52.1%)
Female	116 (47.9%)
Race
Caucasian	130 (53.7%)
African American	104 (43.0%)
Other	9 (3.3%)
Age	58.07 years (12.15)
Marital status
Married	129 (53.3%)
Divorced	49 (20.2%)
Single	37 (15.3%)
Widowed	27 (11.2%)
Education
< High School	49 (20.2%)
High School diploma	67 (27.7%)
Some college	78 (32.4%)
Bachelor’s degree	25 (10.3%)
Post-graduate studies	22 (9.1%)
Employment status
Employed	110 (45.5%)
Unemployed	41 (16.9%)
Retired	70 (28.9%)
Homemaker	5 (2.1%)
Disability	16 (6.6%)
Income
< $10,000	37 (15.3%)
$10K–$29K	60 (24.8%)
$30K–$49K	46 (19.0%)
$50K–$74K	26 (10.7%)
$75K–$100K	32 (13.2%)
> $100K	29 (12.0%)
Declined to answer / Don’t know	12 (5.0%)
Health Insurance
Private	109 (45.0%)
Medicare	68 (28.1%)
Medicaid, State Insurance, Uninsured	65 (26.9%)
Stage
1	19 (7.9%)
2	60 (24.8%)
3	95 (39.9%)
4	60 (24.8%)
Unknown	8 (3.3%)
State
Virginia	167 (69.0%)
Ohio	75 (31.0%)

Associations between sociodemographic characteristics and DD were examined. Gender, race, age, marital status, education, employment status, income, health insurance, and state were not associated with length of DD. However, patients who felt healthcare was unaffordable (n=43) had longer DD (8.8 vs. 4.0 months; p=0.02). Those diagnosed with later stage CRC (Stages 3–4, n=155) had longer DD than those diagnosed with earlier stage CRC (Stages 1–2, n=79) (5.7 vs. 3.3 months; p=0.02).

3.2 Communication in the Consultation: Patient Factors

3.2.1 Healthcare providers consulted

Various health care providers (HCPs) were consulted for initial symptoms. Half of the participants (50.8%, n=123) reported visiting a primary care physician, 14.5% (n=35) went to a hospital emergency room, 5.0% (n=12) consulted a specialist, and 1.7% (n=4) saw an urgent care physician. One quarter (26.4%, n=64) reported visiting more than one HCP about their initial symptoms. Of the 76.4% (n=185) of participants who reported having a regular HCP, 81.0% (n=150) reported consulting with that physician specifically. Having a regular HCP prior to diagnosis (n=185) was associated with longer DD (5.5 vs. 2.7 months; p=0.001).

Follow-up analyses showed that individuals with a regular HCP (n=185) are more likely to seek medical attention faster for initial symptoms (n=79) than those who do not (X²=9.24, p=0.002). Patient charts revealed that those who sought medical attention earlier for initial symptoms (n=79) were more likely to receive a non-cancer diagnosis from the physician (n=77; X²=11.55, p=0.001). Specifically, 37 (15.3%) patients sought medical attention rapidly for initial symptoms but were diagnosed with something other than cancer (misdiagnosis), compared to 120 (49.6%) who only sought medical attention after experiencing symptoms over a lengthy period of time and were not misdiagnosed.

3.2.2 Patient communication styles and strategies

Most patients (n=220, 91.7%) mentioned their symptoms to an HCP, but the number and type of communication styles/strategies used varied. For example, 40.1% (n=97) described using assertive styles of symptom communication. Assertive communication styles included indicating to the HCP that their symptoms were serious (n=68, 28.3%), mentioning a family history of cancer (n=30, 12.5%), asking questions (n=12, 5.0%) asking for tests (n=17, 7.1%), asking if the symptoms could be due to cancer (n=13, 5.4%), or asking for a referral (n=2, 0.8%). More than half (58.7%, n=142) used less assertive communication styles, for example, downplaying symptoms (n=25, 10%), failing to disclose all symptoms (n=38, 16%), mentioning symptoms while discussing other medical issues (n=34, 14%), or offering explanations for symptoms (n=88, 36.7%). A smaller number described being too sick to say very much (n=12, 5.0%) or having trouble describing symptoms (n=5, 2.1%).

3.2.3 Symptom communication barriers

A subset of patients (n=55, 22.9%) described having worries or hesitancy about discussing their symptoms with an HCP. Of those 55, 50.9% (n=28) were apprehensive about the possibility of tests or procedures, 27.2% (n=15) were worried they would be told it was a serious health condition, and 25.4% (n=14) were worried about cancer specifically. Several patients (20.0%, n=11) described feeling embarrassed or did not want to be seen as a “complainer” (12.7%, n=7). Patients who reported difficultly being able to adequately communicate their symptoms to a physician (n=22) had longer DD (9.3 vs. 4.4 months; p=0.001). Patients who described symptoms at a visit scheduled for a reason other than symptoms indicative of CRC (n=61) had longer DD than those who scheduled a specific visit to discuss CRC symptoms (7.4 vs. 4.0 months; p=0.007).

3.3 Communication in the Consultation: Physician Factors

3.3.1 Physician response during the consultation

During the consultation, 45.9% (n=111) of physicians performed a general physical examination, 38.8% (n=94) performed tests and 17.8% (n=43) performed a digital rectal exam. Those patients whose physicians failed to explore symptoms or investigate their complaints (n=78) had longer DD (9.27 vs. 2.78 months; p<0.001). In addition, lower scores on the Relational Communication Task Orientation sub-scale were associated with longer DD (rho=−0.15; p=0.03). The other Relational Communication sub-scale totals were not associated with DD [rho=0.00-(−0.09); p=NS].

3.3.1.1 Physician communication to the patient regarding symptoms

Some physicians adopted communication strategies that moved towards and others adopted strategies that moved away from a CRC diagnosis. Examples of these strategies are listed in Table 2. Patients whose physicians did not engage in communication of next steps (n=73) had longer DD (8.2 vs. 3.4 months; p<0.001). Patients of physicians that engaged in ‘temporizing’ strategies such as telling patients not to worry or to continue to monitor symptoms (n=65), experienced longer DD (6.8 vs. 4.4 months; p=0.006). Those who reported that their physician misunderstood their symptoms (n=17) had longer DD (8.2 vs. 4.7 months; p=0.02), as did those whose physicians did not think the symptoms were serious (n=47; 8.5 vs. 4.0 months; p<0.001). 42 physicians (17.4%) used temporizing strategies and made a clear misdiagnosis compared to 35 (14.5%) who did not use temporizing strategies and made a misdiagnosis (X²=42.91, p<0.001).

Table 2.

Physician communication strategies

Physician communication strategies during visit	n (%)
COMMUNICATION TOWARDS A CRC DIAGNOSIS
Told patient he/she probably had cancer	10 (4.2%)
Gave a list of possible causes for symptoms	24 (10.0%)
Communication of Next Steps	169 (69.8%)
Told patients that tests were needed	135 (56.3%)
Referred to specialist	76 (31.7%)
Explained procedures to be performed	21 (8.8%)
Told patient to see urgent care	10 (4.2%)
COMMUNICATION AWAY FROM A CRC DIAGNOSIS
Gave an alternate diagnosis	53 (22.1%)
Agree with patient’s explanation of symptoms	10 (4.2%)
Temporizing Strategies	65 (26.9%)
Gave advice to improve symptoms (e.g., lifestyle change, meds)	42 (17.5%)
Told patient to return if symptoms do not improve	24 (10.0%)
Told patient to monitor his/her symptoms	11 (4.6%)
Told patient not to worry about his/her symptoms	7 (2.9%)

3.3.2 Consultation outcomes

Half of the sample were either referred to a specialist (26.9%, n=65) or given a prescription to treat symptoms (24.4%, n=59). Others received orders for follow-up tests (18.6% (n=45), had a colonoscopy scheduled for them (18.6%, n=45), were admitted to the hospital (9.9%, n=24), were given a prescription to treat an alternate problem/diagnosis (9.5%, n=23), and/or had a follow-up office visit scheduled (5.8%, n=14). Those whose physicians made a non-cancer diagnosis (n=78) had significantly longer DD (10.2 vs. 2.4 months; p<0.001). Those patients with co-morbid conditions that led to a non-cancer diagnosis (n=40) also had longer DD (9.1 vs. 4.0 months; p<0.001).

3.4 Follow-up and Referral delay

Lack of follow-up included the physician not recommending follow-up at the initial consultation (n=17), the patient not following physician recommendations, such as having a specific test or returning for a follow-up visit (n=21), and/or the physician taking no action to address symptoms at follow-up visit (n=13). Patients with a lack of follow-up (n=30) experienced longer DD (11.1 vs. 3.9 months; p=0.005). Referral delay included the patient not scheduling or showing up at a recommended referral appointment or colonoscopy (n=39), the physician referring the patient to a non-GI specialist (n=11), and/or system factors causing considerable wait time for the referral appointment or colonoscopy (n=27). Those patients with referral delays (n=68) had longer DD (7.5 vs. 3.8 months; p=0.006).

3.5 Factors Associated with DD in Univariate Analyses

Table 3 summarizes the patient and physician consultation communication factors, consultation outcomes, and follow-up and referral delay factors significantly associated with DD.

Table 3.

Summary of consultation characteristics associated with DD

Patient Factors	Mean DD (SD)	t or rho	p value
Having a regular HCP prior to diagnosis**
No (n=56)	2.69 months (8.73)	t = 3.39	0.001
Yes (n=185)	5.51 months (6.32)
Not being able to adequately communicate symptoms to HCP**
No (n=217)	4.44 months (8.00)	t = 3.25	0.001
Yes (n=22)	9.32 months (10.33)
Discussing symptoms at a visit scheduled for another reason**
No (n=179)	4.02 months (7.27)	t = 2.78	0.007
Yes (n=61)	7.35 months (10.56)
Physician Factors
Failure to explore symptoms raised by patient***
No (n=160)	2.78 months (5.60)	t = 6.22	<0.001
Yes (n=78)	9.27 months (11.00)
Relational Communication Task-oriented sub-scale score*		rho = −0.15	0.03
Physician used temporizing strategies**
No (n=200)	4.42 months (8.41)	t = 2.79	0.006
Yes (n=42)	6.83 months (7.52)
Physicians did not engage in communication of next steps***
No (n=169)	3.40 months (6.49)	t = 4.34	<0.001
Yes (n=73)	8.19 months (10.77)
Misunderstands patient’s report of symptoms*
No (n=221)	4.65 months (8.32)	t = 2.41	0.02
Yes (n=17)	8.17 months (8.30)
Does not believe symptoms are serious***
No (n=191)	4.01 months (7.90)	t = 4.00	<0.001
Yes (n=47)	8.52 months (9.21)
Consultation Outcomes
Physician makes a non-cancer diagnosis***
No (n=162)	2.35 months (3.79)	t = 6.99	<0.001
Yes (n=77)	10.21 months (12.04)
Patient has co-morbid conditions that contribute to a non-cancer diagnosis**
No (n=199)	4.04 months (7.44)	t = 3.66	0.001
Yes (n=40)	9.08 months (11.03)
Follow-up and Referral Delay
Overall lack of follow-up**
No (n=212)	3.95 months (7.07)	t = 3.81	0.001
Yes (n=30)	11.14 months (12.74)
Overall referral delay***
No (n=173)	3.84 months (7.59)	t = 3.88	<0.001
Yes (n=68)	7.46 months (9.47)

^*p<0.05,

^**p<0.01,

^***p<0.001

3.6 Predicting diagnostic delay: A logistic regression model

Predictor variables included in the final model were race, education, income, insurance, having a regular HCP, physician temporizing communication strategies, receiving an alternate diagnosis, lack of follow-up, and presence of referral delay. Age, gender and stage of cancer were not significant (p>.10) and not included in the final model. Table 4 displays the results obtained from the logistic regression analysis.

Table 4.

Logistic regression model results

Predictor*	B	S.E.	p-value	Odds ratio
Income	−0.57	0.26	0.03	0.56
Regular physician†	0.93	0.43	0.03	2.52
Alternate diagnosis†	1.21	0.42	0.004	3.36
Temporizing communication strategies†	0.88	0.39	0.02	2.41
Referral delay†	1.28	0.35	<0.001	3.61
Lack of follow-up †	1.20	0.50	0.01	3.32
Education	0.33	0.19	0.08	1.40
Insurance – Private	0.40	0.48	0.40	1.50
Insurance – Medicare	0.13	0.48	0.79	1.14
Race – African American	0.18	0.33	0.60	1.19
R2	0.33	0.07	<0.001

^†Dichotomous variables: Yes=1, No=0

Income was associated with DD (OR = 0.56, p=0.03), such that those participants with lower income were more likely to experience DD >2 months. Patients who had a regular physician prior to receiving a cancer diagnosis were 2.52 (p=0.03) times more likely experience DD >2 months as compared to people who reported not having a regular physician. Individuals whose physicians used temporizing communication strategies during the consultation (OR=2.41, p=0.02), initially gave them a diagnosis other than cancer (OR=3.36, p=0.02), delayed in making a referral to a specialist (OR=3.61, p=<0.001), and patients who experienced follow-up delay of any kind (OR=3.32, p=0.01) all had increased odds of experiencing DD >2 months.

4. Discussion and Conclusion

4.1. Discussion

This paper sheds light on the different strategies patients use in their interactions with physicians about CRC symptoms and physicians’ responses to these communications. Symptoms such as GI distress and constipation are common consultations in general practice and overwhelmingly have causes that are more benign. The challenge for patients is to accurately communicate their symptoms, including the severity and length of duration. For patients with co-morbidities, such as irritable bowel syndrome, it may be difficult to differentiate how these symptoms are different from what they have experienced in the past. For patients whose physicians are familiar with their past gastrointestinal (GI) problems, familiarity may prove a double-edged sword in which a well-known patient’s symptoms may take longer to diagnosis as the physician seeks to convince him or herself that they are different from those experienced in the past. Indeed, with insurers demanding more and more justification for doing expensive tests such as colonoscopy, physicians in the United States are perhaps becoming more cautious in ordering such tests.

These analyses focused on DD with a sample of CRC patients who played a proactive role in recognizing and bringing their symptoms to a physician’s attention. Of all the patient sociodemographic characteristics examined, only income was related to DD, with lower levels of income being associated with longer delay. Unlike other studies, we did not find that race played an independent role apart from income. Moreover, 17.4% of this sample reported having financial issues that caused them to delay seeking care or follow-up care. Other barriers to obtaining a timely diagnosis included faulty initial diagnoses, structural barriers such as physicians delaying a referral for further testing or to a specialist, and patients who failed to follow through with physician recommendation. These findings coincide with those of other studies [12–16].

One communication variable, lack of task orientation, was associated with delay in the final model. Other constructs such as immediacy, i.e., the degree of psychological closeness expressed in a person’s communication with another [33], may have been too distant from the event to accurately report. Immediacy, for example, is cued by vocal warmth, interpersonal touch or reduced physical distance and the interviews were conducted up to 6 months post-diagnosis. This may have affected recall of such events that are frequently noticed and ‘packaged’ within a communicator’s total perception of a communication event.

This was a retrospective study, which is a limitation. In recognition, chart review was used to verify certain aspects of patient self-report. An examination of patients’ medical records revealed that physician charting of symptom documented higher numbers of specific symptoms as compared to reports by patients. Discrepancies were noted especially for GI symptoms such as abdominal pain, diarrhea, constipation, and rectal bleeding. Thus, misdiagnosis, which occurred in 34% of these patients, does not seem to be due to patients not confiding their symptoms to their physicians or retrospectively ‘noting’ symptoms at the interview only. Future analyses will attempt to examine which symptom clusters are more likely to indicate CRC in order to provide greater guidance to primary care physicians making a differential diagnosis.

The importance of effective communication between physicians and patients is highlighted by these analyses. The successful communication approaches include shared decision making strategies such as explaining procedures to be performed and explaining rationales for actions. For example, physicians who planned and communicated next steps to determining the patient’s diagnosis were less likely to have patients who experienced DD. This communication strategy evidently assisted patients’ abilities to take action and obtain further tests that established the diagnosis. Physicians who were not convinced of the severity of the symptoms tended to use temporizing strategies such as asking patients to return if symptoms did not resolve or telling patients not to worry.

4.2. Conclusion

Patient, but especially physician, communication about symptoms and physician ability to distinguish new symptoms from pre-existing GI problems or other non-GI co-morbidities were important factors to prevent DD. In addition, patients who presented with the earliest, less severe symptoms actually experienced longer delay. Finally, lower income was associated with longer DD, an important structural barrier in the context of the US health care system.

The findings of the present study should be interpreted with the following caveats. We did not actually audiotape the conversations between physicians and patients. Reports concerning the quality and content of the conversations are based on patient self-report. Although patients with longer DD may have had a negative recall bias of their communication with their physicians, medical records were reviewed to confirm more objective information such as symptoms reported and visit dates. Agreement between patient self-report and chart review was high. Given the cross-sectional, retrospective nature of this study, causality cannot be inferred.

DD examined in this study does not represent DD of all patients diagnosed with CRC, but a subset of patients who experienced symptoms prior to diagnosis. Results cannot be generalized to the population of CRC patients whose cancers are found on routine screening. The factors predicting delay would likely have been different had these patients been included in this study. Our study’s sample differs from the general population of CRC patients in that it is significantly younger (59 vs. 66 years old), more likely to be diagnosed with Stage 3–4 CRC, and more likely to be African American. In this sample, patients’ appraisals of their symptoms and factors influencing their decision to seek medical care appear to be intertwined with DD. Further research is needed on the relationship between appraisal delay and DD and the factors implicated in each type of delay. Examination of potential mediators (e.g., symptom severity) and moderators (e.g., income) are warranted.

4.3 Practice Implications

This study has implications for how physicians and patients communicate about medical problems during a consultation. Physician-patient discussions about GI symptoms are challenging in primary care practice due to commonalty of the symptoms. Patients sometimes present with new, vague, GI symptoms, some of which may mirror symptoms of existing chronic illness. Thus, careful questioning of the patient is needed to distinguish ‘old’ and similar GI symptoms from newer ones. Furthermore, physicians should be mindful that the first problem the patient discusses may not be the most urgent. Multiple problems presented in a consultation may lead to increased cognitive load (i.e., physicians’ efforts to collect relevant information, maintain the information in working memory, and integrate the information to reach a decision) and ultimately faulty decision making [34].

In addition, when physicians fail to send a clear message about the potential meaning of symptoms or the need for follow-up, they may lull patients into a false ‘sense of security’ interfering with patient actions on recommendations in a timely fashion. This factor, which was labeled as ‘temporizing’, was the strongest communication factor and merits continued exploration. Another recent study of CRC DD also suggests lack of clarity concerning the severity of symptoms may delay obtaining a colonoscopy [17].

Finally, physicians should be alert to the economic challenges their patients may face, which can compromise follow through on future treatment or care recommendations. Providing appropriate resources for obtaining lower cost or more convenient care may help ease financial difficulties and increase patient compliance in a strained economy. In summary, excellent communication skills that appropriately probe for relevant social and economic patient information, assist patients in distinguishing and elaborating on symptoms, foster recognition of symptom seriousness, and provide clear rationale and instructions for future steps will speed along the diagnosis process and could be the difference between early and late stage disease.