Study Design: This descriptive cross-sectional study.
Setting: This study was conducted at the Department of Radiation Oncology, University College Hospital, Ibadan, Nigeria, a 45-bed facility. The department was established in the year 1987 and has Telecobalt and HDR Co60 Brachytherapy machines with modern treatment planning system. The department is a referral centre for cancer patients that require radiotherapy both within the UCH and outside, and as far as some West African countries. From the registry of the department, about 300 new patients with cancer utilize the centre annually [19], thus our sampling frame was taken as 300.
Participants
To be eligible for the study, participants were required to be adult patients, 18 years of age and above with the diagnosis of any cancer. They were consecutively recruited between June 2016 and March 2017. Excluded were patients who were too ill to follow up research protocols, those with a known psychiatric illness, and those patients not willing to give consent. Participants were selected by the use of multi-stage simple random sampling. At the first stage, all the 4 Oncology Departments in the University College Hospital, Ibadan were listed. During the second stage, the Radiation Oncology Department was selected by balloting.
At the Radiation Oncology Department, patients who met the eligibility criteria (n = 130) were consecutively recruited for the study (Figure I). The first subject that was interviewed was selected by simple random sampling and subsequent ones consecutively recruited until the minimum sample was met. An effort was made not to duplicate responses by using their hospital number and later allocating research number on a case by case basis.
Variables
Dependent Variable/Outcome Variable
The dependent/outcome variable was “Any Mental Disorder” (AMD) which was defined as any DSM IV disorder, including delirium and substance use disorder.
Independent Variable
Independent variables were sociodemographic characteristics such as age, sex, employment status, marital status, religion, ethnicity as well as clinical characteristics such as stage of the cancer, duration of treatment, type of treatments received by the patient and presence or medical comorbidity. Other independent variables were presence of psychological distress (GHQ score > 2) and presence of psychosocial distress, DT score > 3).
Confounding Variables
The potential confounding variable was GHQ score > 2 and this was adjusted for during multivariate analysis.
Predictors were variables that were significantly associated with AMD during multivariate analysis, using variables that were significantly associated with AMD during univariate analysis. These were duration of treatment of the cancer, presence of psychosocial distress (DT > 3), and presence of medical comorbidity.
Data Sources
- Sociodemographic and Clinical Questionnaire
The sociodemographic questionnaire yielded information on age, sex, employment of the patient, marital status, religion, ethnicity, stage of the cancer, presence or medical comorbidity, type and number of treatments received by the patient,
- 12-Item General Health Questionnaire (GHQ 12)
The GHQ 12 is a screening instrument for psychiatric morbidity. Although it does not yield a diagnosis, positive scores are indicative of psychological distress. Each item is rated either 0 or 1 on the basis of the frequency with which the subject has experienced the symptom in the recent past, yielding a maximum score of 12. A score of 1 or above is suggestive of psychological distress; however, to increase specificity, a cutoff point of 2 was used, and score of 3 regarded as positive screen [20].
- Distress Thermometer
The Distress Thermometer (DT) was developed as a simple tool to effectively screen for symptoms of distress. The instrument is a self-reported tool using a 0-to-10 rating scale. Additionally, the patient is prompted to identify sources of distress using a Problem List [21].
The cutoff point on the distress thermometer (DT) is 3, and those who score above 3 are classified as having psychosocial distress, which was adopted in the present study. Usually, patients who score a 4 or higher on the DT s more items from the NCCN Problem List [22]. Studies have shown that the DT is positively correlated with the Hamilton Anxiety and Depression Scale [23, 24] and the 12-Item General Health Questionnaire (GHQ 12) [25, 26].
- Delirium
The Confusion Assessment Method (CAM) was used to establish the diagnosis of delirium. The CAM consists of nine questions which correlate with DSM IV criteria for delirium [27]. The questionnaire is administered in about 5 minutes following an initial observation of the patient during a cognitive screening test, the Mini-Mental State Examination (MMSE). The diagnosis of delirium made using a standard algorithm which is part of the CAM instrument. The CAM diagnostic algorithm is based on four cardinal features of delirium: 1) acute onset and fluctuating course, 2) inattention, 3) disorganized thinking, and 4) altered level of consciousness. A diagnosis of delirium, according to the CAM requires the presence of features 1, 2, and either 3 or 4 [28, 29]. The CAM when compared with clinician diagnoses using DSM IV criteria, has a sensitivity of 94–100% and a specificity of 90–95% [30].
- Mini-International Neuropsychiatric Interview (MINI)
The Mini-International Neuropsychiatric Interview (M.I.N.I.) is a brief structured interview designed to make a diagnosis according to DSM-IV and ICD-10 criteria [31]. The instrument has been found to have similar psychometric properties in different parts of the world and different modules of the MINI have been used in the past in Nigerian studies [32-36]. The psychotic, major depressive disorder, bipolar, anxiety, and alcohol and substance modules of the MINI were used to determine the presence or absence of AMD. Adjustment disorder was diagnosed according to the DSM IV-TR criteria by the authors [37].
- Cancer Staging
Cancer staging was performed by one of us (AA) an oncologist and was based on the TNM classification. [38]. According to the classification, Stage 0 is carcinoma in situ for most cancers. This implies the cancer is at a very early stage and has not spread. Example is cervical intraepithelial neoplasm (CIN), Stage I cancer (early stage cancer) means the cancer is small and is confined to one area. Stages 2a, 2b and 3 mean the cancer is larger and has infiltrated the surrounding tissues or lymph nodes, while stage 4 means the cancer has metastasized.
- Medical Comorbidity
Patients were classified as having medical comorbidity by using the proforma used to elicit the presence of medical comorbidity in mental disorders [39]. The proforma included health conditions such as anemia, hypertension, ischemic heart disease, bronchial asthma, chronic dermatitis, diabetes mellitus, thyroid diseases, chronic pain disorder, chronic headache, peptic ulcer disease, liver disease, osteomyelitis, hemorrhoids, tuberculosis, HIV/AIDS, hepatitis schistosomiasis, dracontiasis and malaria. This was supplemented by self- reports of any current health condition with a written medical report by the treating physician.
Treatment Receipt
Treatment receipt was evaluated by asking a yes/no question. “Have you ever received treatment for a mental disorder either in a formal setting such as the hospital or from an informal setting such as the spiritualist in the course of this cancer journey”?
Interviewers
Interviewers were 4 psychiatric resident doctors who had previously received training in clinical research. They had been part of our clinical research time over the course of their training.
Bias
To minimize study bias, a pre-test was conducted before the commencement of the study among 15 patients attending the medical oncology unit of the study centre (not part of current study). This was to ascertain the feasibility, applicability, understandability and the application time of all the instruments of data collection in a busy clinical setting. During this period, it was observed that the research protocols suited the proposed sample and the administration time was about 25 minutes which was regarded as adequate.
Selection bias was removed by employing simple randomly sampling to select the participant. Regression analysis further minimized bias by removing the effects of confounding variables.
Sample Size
Given the population size as 300, at 95% confidence interval, and a 5% margin of error, the minimum sample was calculated using sample size table by Research Advisors as 169 [40]. However, 39 participants were excluded from the study for various reasons, including severe illness, and deaths.
Data Analysis
Data analyses were by descriptive statistics and was carried out on 130 participants. There were no missing data. The GHQ-12 and the DT both have cutoff points. Thus, their association with AMD, the association between sociodemographic characteristics versus AMD, the association between medical comorbidity and AMD as well as the association between cancer stage and AMD were sought using Chi square test. This was because the preliminary Shapiro Wilk test of normality carried out before the final analysis shows that all the independent variables differ significantly from a normal distribution. In order to determine the possible effects of confounding variables, the predictors of AMD were determined using binary logistic regression analysis, using variables that were significantly associated with AMD during univariate analysis. We adjusted for GHQ because of collinearity with DT. All analyses were set at 95% CI and performed using SPSS 20.0.