A quantitative approach using cross-sectional, descriptive-correlational design was used to examine correlates of depression among patients with chronic illness. Data was data from patients diagnosed with diabetes melli- tus-type-II, Rheumatoid arthritis, CAD, cancer, and pulmonary diseases from two major tertiary hospitals in Saudi Arabia. Information collected in regards to depression, stress, and coping.

A convenience sampling of 412 completed and retuned the questionnaire. Sample size was estimated using computer program G. power 3.0 of correlation; normal bivariate model of small effect size of 0.15, at alpha 0.05 two tailed level of significance, and power of 0.8, to be at least 385. A total of 600 patients approached and in- vited to study participation and 412 completed and retuned the questionnaire with a response rate of 69%. The study targeted patients attending primary, secondary and tertiary care units. Inclusion criteria include: 1) diagnosed with one of the following chronic illness longer than 6 months: diabetes mellitus-type-II, Rheumatoid arthritis, CAD, cancer, and pulmonary diseases; 2) age of 18 years or above; 3) ability to read and write in Arabic. Exclusion criteria included: no history of diagnosed mental or cognitive disorders.

Prior data collection, ethical approval obtained from the IRB at King Saud University, and the targeted institu- tions. Data are collected between November 2013 and January 2014. Data collected using self report format of data collection at patient’s convenience. Patients who expressed interest to participation in the study were ap- proached by the researcher who explained the study and provided them with all details and answered all their questions. Patients were asked to sign the consent form that included information related to the title of the study, its purpose, its significance and a statement informing the participants that their privacy would be protected by assuring them that their responses will be treated confidentially, and information that reveal their identity will not be recorded. Also, the information will be used for the purpose of the study, and that their participation is voluntary and they have the right to withdraw at any time during the study and that their decisions will not influ- ence the quality of care they receive. Anonymity of the respondents ensured during and after study completion; and data secured and saved to provide anonymity. Interviews conducted by trained research assistants at private rooms or patients’ rooms upon their convenience. Filling the survey required about 25 minutes and patients who need more time was given opportunity to take break and research assistants retuned to complete filling the ques- tionnaire. The whole package presented in Arabic language.

The data collected using an Arabic version of self-reporting questionnaires. The Instruments were:

1) The Beck Depression Inventory-II (BDI-II) [14] was used to assess patients' depressive symptoms, which contain items that measure cognitive-affective symptoms and attitudes, impaired performance, and somatic symptoms [14] . This instrument contains 21 questions answered on a four-point Likert scale in which 0 represents the absence of symptoms and 3 represents an extreme problem. The total range of 0 to 63 and stan- dard cutoff points as follow: 0 - 13 indicates no or minimal symptom, 14 - 19 indicates mild symptoms, 20 - 28 indicates moderate symptoms, and 29 - 63 indicates severe symptoms [14] . A score of 13 is the cut-off point in- dicating depression. The test-retest r was 0.88, and Cronbach’s Alpha is 0.87 [14] . In this study, Cronbach’s Alpha was 0.79.

2) Stress was measured using the brief form of Psychological Stress Measure [15] . The original Psychological Stress Measure (PSM) was designed using 49 items drawn from descriptors generated by focus groups on stress. The scale is unifactorial in structure and maintains a test-retest stability of 0.68 to 0.80 under apparently constant conditions. Patients checks the answer that best indicates the degree to which each statement has applied to him/her recently The responses made on a Likert scale and ranged from range from 1 (null) to 4 (much). The higher the score in the scale reflect higher level of psychological stress. In this study, Cronbach’s Alpha was 0.81.

3) Coping skills was measured using the abbreviated version of the COPE Inventory [16] . Brief COPE is a 28 items scale measures the ways individuals use to cope with stress in their life. Brief COPE is formed of 14 do- mains (each consisted of 2 items) were responses ranged from 1 (I haven’t been doing this at all) to 4 (I’ve been doing this a lot). The scale takes >10 minutes to be completed. The scale has good internal inconsistency with Cronbach’s alpha of 0.83 [16] . In this study, Cronbach’s Alpha was 0.73.

4) Life satisfaction was measured using the Satisfaction with Life Scale [17] . This is a general measure of life satisfaction, which consisted of five statements. Participants were asked to rate each statement according to the following seven-point scale: 1) strongly disagree, 2) disagree, 3) slightly disagree, 4) neither agree nor disagree, 5) slightly agree, 6) agree, and 7) strongly agree. The scores of the total scale ranges from 5 to 35 and inter- preted as follow: from 31 - 35 (eextremely satisfied), from 26 - 30 (satisfied), from 21 - 25 (slightly satisfied), 20 (neutral), from 15 - 19 (slightly dissatisfied), from 10 - 14 (dissatisfied), and 5 - 9 (extremely dissatisfied). The test-retest reliability was estimated to be 0.87 [17] . In this study, Cronbach’s Alpha was 0.76.

Potential covariates: Gender, age, marital status, type of disease, duration of disease, smoking status, income, education level and work status. The demographic information obtained from an investigator-developed subject profile.

The Statistical Package for Social Science (SPSS 18) software (IBM, Chicago, IL, USA) was considered as suita- ble software for data entry, storage and analysis. Descriptive statistical analysis such as frequency count, percen- tage, mean, median and standard deviation was employed to describe the research sample. The Pearson prod- uct moment correlation coefficient was used to describe the association between variables. T-test for two in- dependent samples (or ANOVA as appropriate) was used to compare means. Statistical significance was set at p > 0.05.

A total number of 412 patients completed the questionnaire (see Table 1). Patients’ age ranged from 18 to 98 years, with mean of 44.2 (SD = 16.8). About 33.3% (n = 138) of the patients there were male patients, while 66.6% (n = 272) were females. In regard to marital status, the majority of them 28.7% (n = 243) were married, while 5.3% (n = 25) were divorced, and 20.8% (n = 85) were single, and 12.8% (n = 53) were widow. The analysis also showed that most of patients (58.5%, n = 242) were not working, and 21.0% (n = 87) of them had a full time work, 13.0% (n = 54) had retired, where the least percent 4.5% (n = 19) of patients had a part time work.

In regard to their medical diagnosis, the analysis showed that 32.5% (n = 134) of the patients had diabetes mellitus type-II, 17.5% (n = 72) had cardiovascular disease, 15.5% (n = 64) had pulmonary diseases, 19.8% (n = 81) had rheumatoid arthritis, and 14.7% (n = 92) had cancer.

P₂₅: percentile 25^th; P₇₅: percentile 75^th.

Depression: Regarding depressive symptoms, the analysis (see Table 2) showed that the patients had a mean score of 14.7 (SD = 10.8) with scores ranging from 0 to 55. About 50% of the patients had a score of 13 or above. In regards to level of depression, the analysis showed that 41.1% (n = 170) of the patients found to have no or minimal depressive symptoms, while 13.8% (n = 57) had mild depressive symptoms, 6.8% (n = 28) had moderate depressive symptoms, and 18.4% (n = 83) had severe depressive symptoms. The analysis indicates that about 45% of the patients are suffering from moderate to severe depressive symptoms compared to 55% with no to mild depressive symptom.

Coping skills: Regarding patients’ coping skills using brief COPE scale (see Table 2), the analysis showed that patients had a mean score of 7.2 (SD = 12.5) with scores ranging from 28 to 110. Considering that the possi- ble range of score is 28 - 112, and that the analysis showed that 50% (n = 206) of the patients had a score of 72 or above and 50% of them had a score between 65 and 80, the results indicate that patients, in general, had moderate ability to effectively cope with their life situations.

Life satisfaction: Regarding patients’ satisfaction about their life (see Table 2), the analysis showed that pa- tients had a mean score of 24.5 (SD = 6.3) with scores ranging from 7 to 35. Considering that the possible range of score is 5 - 35, and that the analysis showed that 50% (n = 206) of the patients had a score of 25 or above and 50% of them had a score between 20 and 29, the results indicate that patients, in general, had high level of satis- faction about their life.

Psychological distress: Regarding patients’ psychological distress level (see Table 2), the analysis showed that patients had a mean score of 38.9 (SD = 11.9) with scores ranging from 12 to 70. Considering that the possi- ble range of score is 9 - 72, and that the analysis showed that that 50% (n = 206) of the patients had a score of 40 or above and 50% of them had a score between 31 and 47, the results indicate that patients, in general, had moderate level of stress.

Using Pearson correlation coefficient (r), the analysis showed that there is a positive and significant correlation between depression and stress (r = 0.33, p < 0.001), and negative a significant correlation with life satisfaction (−0.54, p < 0.001), while no statistical significant correlation found with coping (r = 0.05, p = 0.39). This indi- cates that those who have higher level of stress and lower level of perception of life satisfaction are more likely to have higher level of depression. While, low correlation between coping and depression infers that depression among patients with chronic illness not dependent upon coping strategies used by the patients.

Regarding the relationship between demographic and personal characteristic and depressive symptoms, the analysis showed that there is a significant and negative correlation between patients’ age and depression score (r = 0.17, p ≤ 0.001). Regarding gender differences, the analysis showed that there was no significant difference between male and female patients in their depressive symptoms (t = −0.69, p = 0.488), although mean score of depression for female (15.2, SD = 10.4) patients was higher than males (14.3, SD = 11.5). To examine the dif- ferences in depression in relation to working status, one-way ANOVA was conducted. The analysis showed that there was no significant difference in depressive symptoms in regards to working status, (F_3,325 = 1.65, p = 0.147), and education level (F_3,325 = 1.95, p = 0.071). While there was a significant difference in depression level related to marital status (F_3,325 = 5.82, p = 0.001). Using post hoc comparison (Tukey HSD), the analysis showed

P₂₅: percentile 25^th; P₅₀: percentile 50^th (Median); P₇₅: percentile 75^th.

that single patients had the lowest mean score of depression (M = 10.8) and single patients were the only group of patients whom are statistically significant (low) than all other groups.

Regarding differences in depressive symptoms in relation to patients’ medical diagnoses (see Table 3), the analysis showed that more 24% and 21.2% of those diagnosed with diabetes mellitus and cancer have serve lev- el of depression, while about 10% of the those who have more than one chronic illness suffer severe level of de- pression. Also the analysis showed that about 60% of those diagnosed with arthritis reported no or minimal de- pressive symptoms, however; 32% of them reported moderate to severe level of depression. It’s apparent from the analysis that patients with more than one chronic illness have less depressive symptoms that those who have only one chronic illness. Using ANOVA test to examine the difference in mean score of depression among pa- tients related to their chronic illness, the analysis showed that there were no significant differences (F_5,325 = 1.91, p = 0.09) although there differences in percentages across level of depression related to type of diagnosis.