A Qualitative Study to Assess Collusion and Psychological Distress in Cancer Patients.

INTRODUCTION: Collusion is frequently encountered but least studied entity in palliative care services in India. Impact of collusion is manifold and identifying it requires good communication skills. Once identified, it gives an indication for existing healthy versus developing unhealthy collusion to be dealt within families.
OBJECTIVE: The objective of this study was to identify the prevalence of collusion and its clinical and psychological correlates among patients and caregivers in a palliative cancer care.
MATERIALS AND METHODS: We describe systematic identification and unraveling of collusion across multiple levels in a palliative cancer care eventually drafting an algorithm to unravel the collusion. Patients and families were recruited from in-patient palliative care services after obtaining written informed consent. Qualitative interviews were conducted using collusion questionnaire, EQ5D, Visual Analog Scale, and NIMHANS psychiatric morbidity screen.
RESULTS: Among 62 cancer families interviewed, we identified that 71% collusion exists between doctor and patient, 61.3% between doctor and caregiver, and 75.83% between patient and caregiver. Around 50% collusions were unraveled systematically. Collusion was more prevalent in patients with rapid progression of illness (<6 months), patients with poor coping skills, and preference of being interviewed alone.
CONCLUSION: This statistics suggests that collusion goes unnoticed in terminal illnesses and communication skills play a major role in identifying and dealing with collusion. This also unearths need to formulate interview techniques and structured assessment tools or questionnaire in palliative cancer care which are sparse.

INTRODUCTION

Palliative care is an important component of so-called “end of life care” and is especially important in terminal illnesses like cancer. Pain, fatigue, loss of dignity, depression, and stigma are few disabling aspects for cancer patients. They pose a significant impact on the quality of life of patients as well as their caregivers, collusion is one of them.[1] In simple words, collusion is defined as a secret agreement or cooperation between two or more people who are trying to deceive (Oxford Dictionary. 2nd ed.) In healthcare, collusion implies any information (about the diagnosis, prognosis, and medical details about the person who is ill) being withheld or not shared among individuals involved in care. Collusion also means that relevant and complete medical information is selectively or not disclosed at all to patients and/or relatives.[1] The purpose of this study is to find the prevalence of collusion among cancer patients admitted in a palliative cancer care and their clinical and psychological correlates by systematically asking them a series of questions using collusion questionnaire and address the unnecessary patterns of collusions to demystify and improve the quality of life of patients and caregivers. The aim of this study is to identify the existing communication gaps among patient and family members about diagnosis and prognosis of the cancer and assessing psychological distress associated with it. Earlier studies done in Indian and western setups have shown that collusion is quite prevalent in palliative care services.[2] It is possible that syndromal depression might be absent in terminally ill patients, but constant psychological distress would make them think pessimistically about self, environment, and future. Furthermore, grief about the continuous loss of health may change their perceived disability, spiritual well-being, and expectations from treating team which has been hardly studied in a single study so far. These components are important with respect to changing the quality of life as reported in a study.[3] It is also found that communication gap exists among treating team and patient about not revealing diagnosis or prognosis of an illness, and there is a paucity of standardized interview schedules for this unmet need.[4] This study also highlights a systematic algorithm to interview families involved in unnecessary collusion and to unravel it systematically.

An idea of cure to palliation involves multiple transitions at psychological levels where collusion acts as conspiracy of silence! Based on the individuals involved in direct care of the patient, it can be classified into doctor–patient collusion, doctor–caregiver collusion, patient–caregiver collusion, and doctor–healthcare professional collusion or combination of any of above. Based on the cultural context, collusion is divided into cultural collusion and familial collusion.[5]

Unraveling of unhealthy collusion helps family to focus on pragmatic issues altogether as honest and open communication is less likely to result in untoward consequences. It may help patient to fulfill his unfinished business, repair of relationships, and coming to terms with family and friends. Patients would not regret if false optimism is not maintained (Back AL, 2006). Psychosocial issues are too complicated to handle for untrained staff and having a liaison with psychiatrist encourages proactive dealing with the collusion.

MATERIALS AND METHODS

This is a qualitative cross-sectional study with mixed method design conducted at palliative care ward of Kidwai Memorial Institute of Oncology, Bengaluru, India. The study was conducted for 8 months from September 1, 2015 to April 30, 2016. Ethics committee approved the study and written informed consents were obtained from patients and caregivers. The study was divided into five parts as reflected in a flow chart [Figure 1].

Figure 1

Methodology flow chart

The first part was a case file review of admitted patient selected by simple random sampling allocating random numbers to 10 beds occupied by patients at palliative care ward. Two families were interviewed every week. Odd and even beds in series were chosen every alternate week. This included a detailed understanding of nature of cancer progression, tumor-node-metastasis (TNM) classification, ongoing chemotherapy/radiotherapy/surgery, need for palliation, current analgesic medications, etc., The second part included the assessment of psychological distress among patients which included sociodemographic data sheet, Visual Analog Scale,[6] structured interview to assess psychological well-being, EuroQOL-5D-level-3,[7] and NIMHANS psychiatric morbidity screen (seven items).[8] The interview to assess psychological well-being consisted of items such as distress related to the diagnosis of cancer and acceptance level (anger/acceptance), which was based on the study done earlier.[9] Explanatory model of patient acquiring cancer (denial/bargaining) distress related to adverse effect of medications, Beck's triad of depression, spiritual inclination, faith healing seeking/frequency, coping strategies, most disabling aspect of life – currently and longitudinally, perceived support from spouse/children/parents/other primary caregivers, and expectation from medical professionals (cure/euthanasia). The interview questions were systematically structured by a panel consisting of two psychiatrists and two palliative care specialists. In the third part, qualitative interviews were conducted to identify existing collusion with patient and caregivers together and using collusion questionnaire. It is a 14-item questionnaire to be used in sequential manner where response to previous questions is taken into account for the next question in the series [Figures 2 and 3]. The interview using collusion questionnaire takes approximately 45 min for one interviewee, followed by which a buffer period of 2 days was offered considering high possibility of developing spontaneous communication within family to unravel the collusion naturally. In the fourth part, follow-up interviews with patient and caregiver were carried out on day 3 to watch out for prognosis of collusion. Those families where collusion appeared to start unraveling were encouraged further by a psychiatrist [Figure 4]. The fifth part was prescribing antidepressants and anxiolytics to patients and family members who met the clinical diagnosis (very few patients required it; hence, log of which has not been kept). Patients and caregivers who required medications for the diagnosable depression and anxiety disorders were treated accordingly.

Figure 2

Collusion questionnaire for patients

Figure 3

Collusion questionnaire for caregivers

Figure 4

Algorithm for dealing with an existing collusion

Sample size

Since this study was based on systematic qualitative interviews, saturation to responses was a criterion for determining the sample size. Saturation for selected four items of collusion questionnaire reached at the 60th interview. Considering logistics of the duration of the study in such clinical setting, initially, the sample size was calculated to be 64 based on the following statistics: interviewing two families per week over a span of 32 weeks' (8 months) duration and considering a follow-up interview within 3 days of first interview, the total number of expected families to be interviewed was approximately 64. The final sample size at the end of the study was 62.

Statistics

Frequency distributions of demographic, clinical, and psychological variables were carried out. The Chi-square test was used to determine the goodness of fit with confidence interval of 95% and standard error of 0.05. Binary logistic regression was used to find the correlation of collusion with different parameters using Statistical Package for the Social Sciences (SPSS) version 20 for statistical analysis (IBM SPSS Inc., Armonk, New York 2016).[10]

RESULTS

A total of 62 patients admitted to palliative section of cancer hospital were interviewed for identifying collusion. The mean age of the sample was 50 years (standard deviation = 12.36). It was found that around 60% of them were male and 79% of the patients were married. Eighty-five percentages of patients were belonging to the Hindu religion. Overall, majority of patients were illiterate (75.8%) as compared to professional qualification (6.5%) [Table 1].

Table 1

Sciodemographic profile of patients

	Frequency (%)
Gender
Male	25 (40.3)
Female	37 (59.7)
Marital status
Married	49 (79)
Separated	5 (8.1)
Widowed	8 (12.9)
Religion
Hindu	53 (85.5)
Muslim	8 (12.9)
Others	1 (1.6)
Education
Illiterate	47 (75.8)
Primary	3 (4.8)
Secondary	8 (12.9)
Professional	4 (6.5)
Occupation
Unskilled	56 (90.3)
Semiskilled	5 (8.1)
Skilled	1 (1.6)

Major systems involved in cancer were gastrointestinal, reproductive, and orofacial. Majority of them were TNM Stages 2 or 3 which is why patients required treatment under palliative care services. Eighty-two percentages of patients were already on oral morphine solution for pain control while 66% were undergoing chemotherapy, 50% radiotherapy, and 43.5% underwent recent surgery [Table 2]. However, the systematic records of treatment were not documented in terms of first symptom and first treatment after the diagnosis of cancer as these were not the objectives of the study. Vomiting (27.4%) was the most common adverse effect found among patients with palliative needs which may be due to the adverse effect of chemoradiotherapy and opioids; interestingly, 64.5% were free of any adverse effects [Table 2].

Table 2

Cancer profile of patients

	Frequency (%)
Cancer affected system
Gastrointestinal and hepatobiliary	16 (25.8)
Reproductive	14 (22.6)
Orofacial	13 (21)
Blood cells	5 (8.1)
Breast	5 (8.1)
Respiratory	4 (6.5)
Other system involvement	5 (8.1)
TNM stages of cancer
2	27 (43.5)
3	28 (45.2)
4	7 (11.3)
Treatment given to the patients
Chemotherapy	41 (66.1)
Radiotherapy	31 (50)
Surgery	27 (43.5)
Oral morphine solution for pain control	51 (82.3)
Adverse effect of treatment
Nil	40 (64.5)
Vomiting	17 (27.4)
Lymphedema	5 (8.1)

TNM: Tumor node metastasis

Regarding psychological needs of patients, it was found that 56% had healthy acceptance of the illness as per Kubler-Ross stages of assessment of grief. There was a significant correlation of development of collusion with poor coping skills, high psychological distress, and rapid progression of illness, that is, shorter duration of illness [Tables 3 and 4]. Thirty-seven percentages of patients attributed the illness to God following attributions to karma (16%), medical reasons (27%), and personalized model of illness (19.4%) as explained through few examples in Table 5. Surprisingly, only 1.6% of patients were found to have unhealthy coping associated with illness and its progression, and only 14.5% of patients reported to have pessimistic view of self, environment, and future while 85% did not. Nearly 67% of patients had clear understanding that illness has nothing to do with faith healing or other forms of treatments while it was found that patients' inclination toward spirituality was very high (>70%) as compared to absent spiritual support (3.2%)

Table 3

Psychological profile of patients interviewed for collusion (n=62)

Frequency (%)
Kubler-Ross stages of coping
Acceptance	35 (56.5)
Anger	8 (12.9)
Bargaining	11 (17.7)
Denial	8 (12.9)
Explanatory model of patient in acquiring the illness
Attribution to God	23 (37.1)
Attribution to Karma	10 (16.1)
Medical reasons	17 (27.4)
Personalized model	12 (19.4)
Coping with the illness
Poor	1 (1.6)
Not so bad	14 (22.6)
Good	26 (41.9)
Excellent	21 (33.9)
Beck's triad of depression
Yes	9 (14.5)
No	53 (85.5)
Spiritual inclination
Absent	2 (3.2)
Mild to moderate	15 (24.2)
Moderate to high	29 (46.8)
High	16 (25.8)
Faith healing/magico-religious treatment seekingtendencies
Absent	18 (29)
Low	24 (38.7)
Medium	13 (21)
High	7 (11.3)
Disabling aspect of life at present situation
Pain	46 (74.2)
Dysphagia	4 (6.5)
Weakness	3 (4.8)
Bleeding	2 (3.2)
Dyspnea	2 (3.2)
Disfigurement	1 (1.6)
Dysarthria	1 (1.6)
Stigma	2 (3.2)
Micturition frequency	1 (1.6)
Perceived support
Spouse	24 (38.7)
Children	12 (19.4)
Both	19 (30.6)
Others	7 (12.3)
Expectation from health professionals
Cure	35 (56.5)
Palliation	20 (32.3)
Death/euthanasia	5 (8.1)
Others	2 (3.2)
Disabling aspect of life in future
Family worries	23 (37.1)
Pain	17 (27.4)
Work/job	12 (19.4)
Basic needs	2 (3.2)
Weakness	1 (1.6)
Illness	1 (1.6)
Nil	6 (9.7)

Table 4

Correlation of collusion with statistically significant variables

Variable	SE	P	CI
			Lower limit	Upper limit
Duration <6 months	0.742	0.008	0.033	0.598
High level of psychological distress	0.814	0.014	0.027	0.667
Preference for interviewing alone by patient	0.753	<0.001	0.012	0.224
Readiness of patient to discuss with family members	1.136	0.004	3.296	9.534
Readiness of patient to inform details of illness	1.120	0.002	3.626	29.126

SE: Standard error, CI: Confidence interval

Table 5

Case examples illness models in cancer patients

1. A 70-year-old Muslim gentleman was angry toward his wife when admitted to cancer hospital after shifting from the general hospital where diagnosis revealed high-grade squamous cell carcinoma of the lung with pleural effusion. The patient denied that he had any illness ever, instead persecuted his first wife and a sister for doing black magic on him and giving him such pains

2. A 55-year-old Hindu gentleman after diagnosing secondary metastasis at the lower end of humerus underwent resection of bone metastasis, followed by abdominal pain. In that intense agony, he had not forgotten to curse his previous surgeon who operated on bone metastasis. Somehow, he was sure that doctor had done something wrong with him which is causing abdominal pain now. He investigated across multiple hospitals and finally visited the one where we were seeing him. He broke down when spoke about eight multispecialty hospital journey still being unaware of his diagnosis and still hoping to get discharged soon and visit another multispecialty hospital. This gentleman was counseled and relieved over next 1 week

3. A 50-year-old truck driver presented with recurrence of multiple myeloma after 5 years. While crying he stops and talk about symbolic cure connected 5 years back when a specific honey bee sting coincided with his admission in hospital for an intense agonizing pain in back and was found to have bone metastasis at L5S1. He recovered through chemoradiotherapy and surgery. This time, the symbolic help and symbolic cure can only be brought back through symbolic repeated honey bee stings and current admission to cancer hospital for relapse of myeloma. He denied that he has any serious illness

The most disabling aspect of cancer was pain according to 74% of patients, followed by dysphagia, dyspnea, and disfigurement. The most common perceived support for a cancer patient was his or her spouse (38.7%), followed by children (19.4%) and others (12.3%). The most common expectation of a cancer patient from his treating physician was cure (56%), followed by palliation (32.3%) and death or passive euthanasia (8.1%). When they were asked about their future concerns, the most common response was family-related worries (37.1%), followed by pain relief (27.4%) and occupation-related issues [Table 3]. Around 10% of patients did not report any specific concerns about their future. Summary of interview preferences and common responses from patients and their caregivers about readiness to unravel the collusion are described in respective tables [Tables 6, 7a, 7b, 8a and 8b].

Table 6

Interview preferences given by patient for collusion interview (n=62)

Interview preference given by the patient	Frequency (%)
Alone	42 (67.7)
In presence of caregiver	13 (21)
In presence of caregiver but negative	7 (11.3)
gesturing by relatives

Table 7a

Caregiver's responses to collusion questions from the most common 1 to least common 10

Caregiver's response	What would happen if patient come to know about the condition
1	Will break down
2	Will be sad and depressed
3	Will not cooperate for treatment
4	Will ask to take treatment from other hospitals
5	Will think of dying
6	Will stop taking food and water
7	Will ask for discharge immediately
8	Will be angry and scold us
9	Will be afraid of touching us and sharing food
10	Will make his illness progression fast due to tension

Table 7b

Caregiver's responses to collusion questions from most common (1) to least common (8)

Caregiver's response	What would happen if the patient do not come to know about the condition
1	Will have good quality of life
2	Will be able to live longer
3	I do not know
4	Will stay happy
5	Will continue to accept the treatment
6	Will continue to worry about what is happening with his/her health
7	Will stop using tobacco at least
8	You tell me doctor, what to do?

In terms of interviewing patient on collusion questionnaire when preference was asked, 67.7% preferred to talk alone with therapist while 32% preferred with relatives, out of which 11.3% were supported by relatives through answering certain uncomfortable questions and gesturing to physician to not disclose certain issues in front of patients as depicted in Figure 5. This was statistically significant with existence of collusion in a family [Table 4]. Some of the common responses used in response to collusion questionnaire by patient as well as caregivers are presented in Figure 6. Around 82.3% of the patients and 48.4% of caregivers reported that they did not know about entity cancer before they reached current hospital. There was a significant correlation of patient's unwillingness to discuss with relatives and existence of collusion in a family [Table 4].

Figure 5

Interview preference by patients

Figure 6

Common responses to collusion questionnaire

It was found that unwillingness to communicate about the illness was highly prevalent among all three groups with figure of 71% between doctor and patient and 61.3% between doctor and caregivers while 75.83% between patient and caregiver depicted in Figure 7. With work that was carried out intended to support family in terms of their psychological needs and to improve the quality of life, almost 41.9% of doctor–patient and doctor–caregiver collusions were unraveled while 33.9% were left untouched as shown in Figure 8.

Figure 7

Percentage of collusion between three groups

Figure 8

Percentage of collusions unraveled

DISCUSSION

Cancer is a multifactorial disease, and its management is multidimensional which involves chemotherapy, radiotherapy, surgery, and palliation. In patient's understanding, this is a very complex management and communicating various aspects of cancer, and its treatment can help substantially to halt the collusion development. More than 75% of the patients interviewed in the study were illiterate and involved in unskilled occupation. This may be due to a reason that study was conducted in a tertiary care government institute which provides subsidized rates of treatment for poor patients and also being one of the largest cancer care hospitals with palliative care services in the region. Hence, it is highly recommended to impart them basic understanding of various illness parameters. Hindu, married and male patients were overrepresented in this study due to methodological limitations. It is possible as Hindu community is predominant population of the region, may be more open to medical model of illness than others. The substance use has been associated with an increased incidence of cancer.[11] In the current study, around 66% of the patients had exposure to tobacco and alcohol before the development of cancer. Collusion is very common in palliative cancer care services (76%) in current clinical observation.[12] Stepwise communication toward patient and caregiver can break unhealthy collusion (42% in the current presentation).[13] Collusion can interfere with treatment seeking, compliance, and interpersonal family dynamics. Handling collusion effectively may help to improve the overall quality of life. More than 50% of families attributed illness to supernatural powers such as “God” and “Karma” possibly a result of culturally acceptable definitions of sufferings while attribution to medical reasons indicates high level of understanding about illness through witnessing such patients or watching advertisements in media. Personalized models of an illnesses stems from denial about the illness since most of them reflect curable etiological models as explained earlier in Table 5.[1] Collusion was highly correlated with high psychological distress and poor coping skills among patients which might explain the caregiver's goodwill to protect them by not revealing the truth in some cases. The study findings also reflect that clinical depression is less common than psychological distress and anxiety among patients involved in palliative care.[14] Despite which more than half of the patients and family members were aware that they had a serious illness and it has no cure but did not discuss about it with family members which indicates that it is worthwhile to talk about “end of life issues,” dying process and unfinished works with them rather that keeping it a secret. The most common disabling aspect in terminal cancer is pain[15] that was consistent finding with this study. However, prospectively, family worries were more disabling for patient and pain ranked second in the list. Around 68% of patients preferred to be interviewed in the absence of caregiver, and this factor was statistically significant with existence of collusion in family. This clearly open up the discussion whether we as a clinician are not adequately equipped with communication skills and further it brings up a question that quality of life of patient and family could have been better if communication gap between family and patient is effectively reduced by dealing with collusion though no statistical correlation was found between collusion and quality of life in this study. Figure 4 reflects the workflow to unravel the collusion systematically. To identify and deal with the collusion is a challenging task,[16] which requires systematic approach to screen and develop innovative psychological intervention to improve the service delivery in palliative care.[13] We need to deliver the message that even the worst disorders have the potential for an unexpected outcome. In serious illnesses such as cancer, providing numbers is like a death sentence (e.g., median survival time). It is better to not give false optimism and screen for psychological distress more than syndromal diagnosis.[5] Each clinician can formulate words and phrases that capture paradox of uncertainty to sustain hope and honesty which might require certain level of communication skills in breaking bad news and dealing with collusion. After all, unraveling of the collusion can also help us to divert focus from body to soul and fits well in definition of healing in the Indian context.[1] Reading patient's needs at particular stage of serious illnesses is a key to deal with treatment refusal and treating noncompliance. Sometimes, narrative catharsis helps for self-healing as we are aware that medical uncertainties are certain!

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Table 8a

Patient's responses to collusion questions from most common (1) to least common (5)

Patient's response	What would happen if family members come to know about the condition
1	Will be upset and sad
2	Will become emotional and cry
3	Will be burdened by the fact that I cannot support them anymore
4	Will be in terrible condition and fearful
5	Will worsen their health conditions

Table 8b

Patient's responses to collusion questions from most common (1) to least common (3)

Patient's response	What would happen if family members do not come to know about the condition
1	They will never believe that I have cancer
2	They probably know partially/completely about my cancer
3	They will take care of me