|Year : 2022 | Volume
| Issue : 2 | Page : 101-106
Psychometric properties of Kessler's Psychological Distress Scale (K10) in cancer patients
Manish Namdeo Thakre1, Harshal Shriram Sathe2, Manoj Rajanna Talapalliwar3
1 Associate Professor, Department of Psychiatry, Government Medical College, Nagpur, Maharashtra, India
2 Assistant Professor, Department of Psychiatry, Mahatma Gandhi Institute of Medical Sciences, Sevagram, Maharashtra, India
3 Associate Professor, Department of Community Medicine, Government Medical College, Gondia, Maharashtra, India
|Date of Submission||15-Jul-2021|
|Date of Acceptance||09-Jan-2022|
|Date of Web Publication||08-Apr-2022|
Dr. Harshal Shriram Sathe
Department of Psychiatry, Mahatma Gandhi Institute of Medical Sciences, Sevagram, Wardha - 442 102, Maharashtra
Source of Support: None, Conflict of Interest: None
Context: Kessler's Psychological Distress Scale (K10) is a self-rated, easy-to-apply scale to measure psychological distress. The suitability of its use in treatment-seeking cancer patients in the clinical setup has not been studied.
Aims: The aim was to assess the psychometric properties and associations of K10 with sociodemographic and clinical variables and depression levels in the cancer patients visiting the hospital.
Settings and Design: The research was a cross-sectional, observational, descriptive study conducted in the oncology outpatient department of a public hospital.
Materials and Methods: The data were collected from 155 cancer patients using semi-structured pro forma for sociodemographic and illness-related information. K10 and Patient Health Questionnaire-9 were used to quantify psychological distress and depression.
Statistical Analysis Used: The factor structure of Kessler's 10-item scale was assessed by confirmatory factor analysis, and Cronbach's alpha was calculated as a measure of internal consistency. Given the nonnormal distribution of quantitative data, nonparametric tests were used to analyze the association of K10 scores with sociodemographic and clinical variables and depression scores.
Results: The K10 showed good internal consistency (Cronbach's alpha = 0.914) in the cancer patients. In confirmatory factor analysis using structural equation modeling, the single-factor and two-factor models could not adequately fit across goodness-of-fit indices. There was a significant association between the levels of psychological distress and depression in cancer patients.
Conclusion: K10, in its current form, is a reliable instrument to measure psychological distress. However, a need-based modification of the existing scale is required in treatment-seeking cancer patients.
Keywords: Cancer, factor structure, K10, psychological distress, psychometric properties
|How to cite this article:|
Thakre MN, Sathe HS, Talapalliwar MR. Psychometric properties of Kessler's Psychological Distress Scale (K10) in cancer patients. Arch Ment Health 2022;23:101-6
|How to cite this URL:|
Thakre MN, Sathe HS, Talapalliwar MR. Psychometric properties of Kessler's Psychological Distress Scale (K10) in cancer patients. Arch Ment Health [serial online] 2022 [cited 2023 Mar 31];23:101-6. Available from: https://www.amhonline.org/text.asp?2022/23/2/101/342749
| Introduction|| |
Psychological distress is an unpleasant psychological response of an individual to internal or external stresses. People having cancer often experience high levels of unrecognized psychological distress. This distress arising from various factors, such as pain, functional disability, and imminent fatality, often goes unaddressed. The presence of psychological distress negatively affects the coping ability, pain perception, and treatment of cancer patients. Similarly, the mental health disorders recognized by standard classification systems such as depression and anxiety are highly prevalent in cancer patients and negatively affect their quality of life. Given the high prevalence of mental health problems and their adverse impact on illness course and management, routine screening of cancer patients for these problems is desirable. A short and easy-to-use psychometric screening tool that measures psychological distress and significantly predicts depression and anxiety can be valuable in the prompt identification of these mental health problems in busy public hospital cancer clinics.
Kessler's Psychological Distress Scale (K10), having 10 items, was originally designed for measuring distress in population samples. The psychometric properties of K-10 have been tested in various populations, and it is a valid and reliable risk assessment tool for psychiatric disorders in different cultures and settings. An Indian study done in the general population of Goa state concluded that K10 has high accuracy and internal consistency in the diagnosis of common mental disorders. Accordingly, the scale has been used in field studies in India to assess mental health issues. However, researchers are coming up with recommendations to use the scale in clinical settings in various patient groups. It has been successfully tested in clinical setups on pregnant and postpartum women, as well as intravenous drug use patients., The utility of the K10 scale in terms of validity and reliability has never been tested in cancer patients, despite their vulnerability for psychological problems.
The developers of K10 have suggested using this scale as a single-factor assessment tool where all the 10 items are loaded into one factor of psychological distress. However, subsequent researches proposed at least four different models which fit the study population better than the single-factor model. The psychometric properties of K10 in the patients attending clinics have been assessed in very few previous studies. Hence, the current study was done to assess the reliability and factor structure of the K10 scale in the cancer patients visiting the hospital. We also explored the association of psychological distress with the demographic and clinical variables and the level of depression in these patients.
| Materials and Methods|| |
The study began after the approval from the institutional ethics committee in the outpatient department (OPD) of the oncology unit at a tertiary health-care center and a teaching hospital in Central India. The study had a cross-sectional observational descriptive study done on the population of treatment-seeking cancer patients visiting the hospital. A trained psychiatric social worker was recruited for data collection who would visit the cancer OPD in morning hours, thrice in a week. Every tenth patient who got registered in the OPD for cancer treatment was recruited in the study after taking informed consent. The process of data collection was periodically monitored by the investigators. The sample of 159 male and female cancer patients was collected in a 3-month period. A thorough assessment of case record forms was done for missing data. Four records were excluded because of missing data, and the data of 155 subjects were analyzed. As per the globally accepted rule of thumb for sample size calculation in confirmatory factor analysis, the minimum sample required is 10 times the total number of items on the scale. As K10 has ten items, the minimum sample size required for the current study would be 100. Hence, our sample size of 155 exceeded the minimum requirements for the test. Sociodemographic and cancer-related information of the patients was collected on a semi-structured pro forma. Psychological distress was measured using Kessler's Psychological Distress Scale (K10), and the Patient Health Questionnaire-9 (PHQ-9) was used to quantify depressive symptoms.
The data were collected using a predesigned pro forma. The sociodemographic and cancer-related information of the participants was collected in the first section. In the second and third sections, the participants answered the Hindi language versions of K10 and PHQ-9, respectively.,
Kessler's Psychological Distress Scale (K10)
The participants were asked to rate ten items on this five-point Likert scale based on their experience of psychological symptoms for the past 4 weeks. Each item of scale had a minimum score of 1 if the participant marked “none of the time” and a maximum score of 5 if the symptom was marked to be present all the time. Thus, the range of total score of K-10 was 10–50, with higher scores showing higher levels of psychological distress. The scale does a dimensional assessment of the construct of psychological distress and does not have well-established cutoff scores for categorical analysis.
Patient Health Questionnaire-9
The PHQ-9 is a validated instrument used to make a criteria-based diagnosis of depression in primary care patients. It has a shorter length but comparable sensitivity and specificity to other commonly used instruments to measure depression. This scale records patients' experience of various depression symptoms over the past 2 weeks on a four-point ordinal scale. The lowest response for the items is “not at all” and the highest response is “nearly every day” based on the frequency of symptoms experienced by the patient. The patients are categorized into the groups of major depressive syndrome, other depressive syndromes, and those not having depression based on the responses on the scale. Each item on the scale was also scored in the range of 0 (for the response as not at all) to 3 (for the response of nearly every day), and the total scale scores were used for association and correlation statistics.
The data were entered into the EPI INFO 2007 Software, Centre for Disease Control and Prevention (CDC).Atlanta, Georgia, United States of America, and data cleaning was done. Finally, 155 records were transferred to JASP Software, Department of Psychological Methods, University of Amsterdam, Amsterdam, Netherlands. We used the JASP software for descriptive and inferential statistics and confirmatory factor analysis. The categorical variables were expressed in numbers and percentages. All the quantitative variables were tested for normality. Nonparametric tests were used for variables that were not following a normal distribution.
The factor structure of Kessler's 10-item scale was assessed by confirmatory factor analysis using structural equation modeling. We examined two models, a single-factor model and the two-factor model. The single-factor model had all 10 items loading on a single factor representing general psychological distress, as suggested by Kessler et al. In the two-factor model, the items loaded on two correlated first-order factors, namely depression and anxiety, as suggested by Sampasa-Kanyinga et al. in their research on the Canadian military population [Figure 1] and [Figure 2]. For analysis, each item was considered independent. Cronbach's alpha was calculated as a measurement of composite reliability or internal consistency. The value of Cronbach's alpha above 0.7 was considered good internal consistency. The goodness of fit was examined by Chi-square test, comparative fit index (CFI), Tucker–Lewis index (TLI), goodness-of-fit index (GFI), and root mean square error of approximation (RMSEA). Model fit was contemplated as acceptable if the null hypothesis was accepted in Chi-square goodness-of-fit test and for the GFI more than 0.90, CFI and TLI more than 0.95, and RMSEA <0.08.
|Figure 1: Confirmatory factor analysis of single-factor model of psychological distress. DIS: Psychological distress, K10x: Items of K10 with x denoting order number in original version of scale|
Click here to view
|Figure 2: Confirmatory factor analysis of two-factor model of psychological distress. DEP: Depression, ANX: Anxiety, K10x: Items of K10 with x denoting order number in original version of scale|
Click here to view
The associations between categorical variables and ordinal variables not following normal distribution were tested by Mann–Whitney U and Kruskal–Wallis test. Spearman rank-order correlation analysis was used to study the correlation of quantitative and ordinal variables such as age, education level with the scale scores of depression, and psychological distress. The correlation between scores of K10 and PHQ-9 was also done using the Spearman correlation coefficient.
| Results|| |
Demographic characteristics of study participants
The mean age of the study participants was 50.45 (±11.53). Majority of the total study participants (n = 155) were females (n = 109, 70.3%). The level of education was above secondary school education in 91 respondents (58.7%). Most of the study subjects were married (n = 145, 95.5%). Almost equal number of the participants lived in rural (n = 75, 48.4%) and urban (n = 80, 51.6%) areas. Gynecological and breast malignancies were the most common cancer groups found in our study subjects (n = 89, 57.4%), followed by gastrointestinal cancers (n = 40, 25.8%). Eleven (7.1%) subjects had a past history of mental illnesses, whereas 30 (19.3%) participants had known physical comorbidities such as diabetes and hypertension [findings represented in [Table 1]].
Internal consistency and confirmatory factor analysis
The K10 scale showed good internal consistency after including all 10 items (point estimate of Cronbach's alpha = 0.914). In confirmatory factor analysis using structural equation modeling, the single-factor and two-factor models could not fit adequately across the goodness-of-fit indices. Specifically, the single-factor congeneric model proposed by Kessler et al. did not adequately fit through our sample as there was a significant difference in the Chi-square goodness-of-fit test (P = 2.95e−11, χ2 = 120.20, dF = 35). The other indices also could not adequately accept the model (CFI = 0.902, TLI = 0.874, GFI = 0.858, RMSEA = 0.126). We repeated the analysis after removing the fifth item (“how often did you feel restless and fidgety”) and the seventh item (“how often did you feel depressed”) which had loading estimates of <0.5. However, removal of these items did not improve model performance (P = 5.64e−10, χ2 = 84.925, dF = 20) or goodness-of-fit indices (CFI = 0.910, TLI = 0.873, GFI = 0.873, RMSEA = 0.146).
The analysis for the two-factor model as used by Sampasa-Kanyinga et al. also showed an inadequately fit model (Chi-square goodness of fit, P < 0.001, χ2 = 298.420, dF = 35; CFI = 0.697, TLI = 0.611, GFI = 0.802, RMSEA = 0.222). We repeated the analysis after removing the poor loading fifth and seventh items for model improvement. The model performance did not improve on the Chi-square test (P < 0.001, χ2 = 247.472, dF = 35) and other goodness-of-fit indices (CFI = 0.683, TLI = 0.557, GFI = 0.806, RMSEA = 0.276) after the modification [Table 2]. Thus, the two-factor model was an even poorer fit than the single-factor model [findings represented in [Figure 1], [Figure 2] and [Table 2].
Magnitude of psychological distress and depression and their association with sociodemographic characteristics in study subjects participants
The median score (interquartile range) of the K10 scale and PHQ-9 scale in the study participants was 28.00 (22.5–32.0) and 8 (5–10). K10 score was significantly associated with marital status (P = 0.002) and showed a positive correlation with increasing age (P = 0.001) and mean PHQ-9 score (P = 1.34e−25). PHQ-9 score also had a significant association with marital status (P = 0.013). The levels of psychological distress and depression did not vary significantly among patients with cancers involving different organ systems. On using the diagnostic criterion for depression recommended by the authors of this scale, 27.7% of participants were found to have “other depressive syndromes” whereas 4.5% had “major depressive syndrome” [findings represented in [Table 3]].
|Table 3: Associations and correlations of psychological distress and depression|
Click here to view
| Discussion|| |
K10 has a proven utility to measure distress in the general population, but its usefulness in the clinical patient samples has not yet been adequately studied. Earlier research shows it has high levels of internal consistency and concurrent validity in diagnosing depression and anxiety in substance use disorder patients. These findings were replicated in our study in the cancer patients where K10 showed high levels of internal consistency (Cronbach's alpha = 0.914). Thus, all the items of scale show a good amount of inter-relatedness in measuring the construct of psychological distress. High levels of alpha show unidimensionality or homogeneity, which is consistent with the developers' proposal of a single-factor model of the scale. This finding echoed in the current study where the goodness-of-fit indices for the single-factor model were better than those for the two-factor model.
K10 is a nonspecific measure of psychological distress developed as per the item response theory. A previous research into the factor structure of K-10 in population-based samples has produced confusing results. K10 was originally proposed by the developers for dimensional measurement of a single factor of psychological distress. However, Brooks et al. contrasted this with a multifactorial model where items loaded in four first-order factors (nervousness, agitation, fatigue, and negative affect) which further loaded in the 2 second-order factors of depression and anxiety. Thus, there is no clear consensus on the exact factor structure of K10. The confirmatory factor analysis of K10 in the treatment-seeking anxiety patients in clinical settings found an absence of fit for the single and a multifactorial model. Similar results were found in the present study on cancer patients which investigated the factor structure of K10 regarding the proposed single-factor and two-factor models. The items on the K10 scale are limited to the evaluation of fatigue, nervousness, agitation, and depressed mood. Hence, the poor fitting of the single-factor model may be explained by the absence of items evaluating pain, functional disability, or unwanted thought ruminations which often underlie psychological distress in cancer patients.,, Although the scale is known to screen depression in primary care patients, it did not fit the model with depression and anxiety as factors. This may be because the scale lacks in the items measuring symptoms commonly associated with depression (changes in sleep and appetite, cognitive decline, and suicidal ideation) and anxiety (motor tension and autonomic hyperactivity) in cancer patients. Hence including more variables may make the scale suitable for assessing distress in cancer patients.
K10 scale has certain disadvantages, such as the ability to evaluate a limited number of factors (depression and anxiety), leaving other important factors such as psychosis and suicidal ideation unevaluated. We tested no more models of K10 in the present study to prevent a chance finding of a fit model that does not reflect the latent true scale structure. Although K10 has been found useful to screen mental disorders in the clinical settings in patients of intravenous drug abuse and pregnant women, our findings imply that the existing variable structure of K10 needs to be revised for patients with chronic physical illnesses such as cancer.,
One of the major strengths of the present study was to assess the factor structure of K10 in the previously unexplored population of hospital visiting cancer patients, who are vulnerable to developing psychological distress. The previous studies have explored the utility of this scale in population-based samples. Our analysis upholds the developer's version of the single-factor model of K10 to measure psychological distress as the goodness of fit for the single-factor model was found better than the two-factor model which measures depression and anxiety. Hence, the present research highlights the difference between the construct of psychological distress, which develops as a response to stressful situations (suffering from a terminal medical illness), and constructs of depression and anxiety, which are multifactorial in origin., However, the present work was limited to the assessment of factor structure of K10 scale in its existing form, and the modification in the items or addition of newer items to improve the goodness of fit was not studied.
| Conclusion|| |
The K10 scale has been developed to measure psychological distress in population surveys. However, the ease of application and availability of this instrument in local languages may prompt its use in the patients visiting the hospital for treatment. The findings of our research challenge the suitability of using K10 in this patient population. Further research is warranted to make a need-based modification of the existing scale to improve its utility in cancer patients. The factor structure of K10 can also be studied in patients having other chronic medical illnesses such as diabetes, heart diseases, and multiple sclerosis to assess its applicability as a screening tool in these conditions. The extension of the K10 scale to include more items for the better measurement of depression and anxiety may also be pursued by future researchers.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Simard M, Hudon C, van Reekum R. Psychological distress and risk for dementia. Curr Psychiatry Rep 2009;11:41-7.
Ryan H, Schofield P, Cockburn J, Butow P, Tattersall M, Turner J, et al.
How to recognize and manage psychological distress in cancer patients. Eur J Cancer Care (Engl) 2005;14:7-15.
Hejazi F, Bahrami M, Keshvari M, Alavi M. The effect of a communicational program on psychological distress in the elderly suffering from cancer. Iran J Nurs Midwifery Res 2017;22:201-7.
Alagizy HA, Soltan MR, Soliman SS, Hegazy NN, Gohar SF. Anxiety, depression and perceived stress among breast cancer patients: Single institute experience. Middle East Curr Psychiatry 2020;27:1-0.
Kessler RC, Andrews G, Colpe LJ, Hiripi E, Mroczek DK, Normand SL, et al.
Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med 2002;32:959-76.
Fassaert T, De Wit MA, Tuinebreijer WC, Wouters H, Verhoeff AP, Beekman AT, et al.
Psychometric properties of an interviewer-administered version of the Kessler Psychological Distress scale (K10) among Dutch, Moroccan and Turkish respondents. Int J Methods Psychiatr Res 2009;18:159-68.
Patel V, Araya R, Chowdhary N, King M, Kirkwood B, Nayak S, et al.
Detecting common mental disorders in primary care in India: A comparison of five screening questionnaires. Psychol Med 2008;38:221-8.
Sathe HS, Mishra KK, Saraf AS, John S. A cross-sectional study of psychological distress and fear of COVID-19 in the general population of India during lockdown. Ann Indian Psychiatry 2020;4:181. [Full text]
Spies G, Stein DJ, Roos A, Faure SC, Mostert J, Seedat S, et al.
Validity of the Kessler 10 (K-10) in detecting DSM-IV defined mood and anxiety disorders among pregnant women. Arch Womens Ment Health 2009;12:69-74.
Hides L, Lubman DI, Devlin H, Cotton S, Aitken C, Gibbie T, et al.
Reliability and validity of the Kessler 10 and Patient Health Questionnaire among injecting drug users. Aust N Z J Psychiatry 2007;41:166-8.
Blanc S, Zamorski M, Ivey G, McCuaig Edge H, Hill K. How much distress is too much on deployed operations? Validation of the Kessler Psychological Distress Scale (K10) for application in military operational settings. Mil Psychol 2014;26:88-100.
Berle D, Starcevic V, Milicevic D, Moses K, Hannan A, Sammut P, et al.
The factor structure of the Kessler-10 questionnaire in a treatment-seeking sample. J Nerv Ment Dis 2010;198:660-4.
Myers ND, Ahn S, Jin Y. Sample size and power estimates for a confirmatory factor analytic model in exercise and sport: A Monte Carlo approach. Res Q Exerc Sport 2011;82:412-23.
Choo K, Spitzer RL, Williams JB. The PHQ-9. J Gen Intern Med 2001;16:606-13.
Dean AG, Arner TG, Sunki GG, Friedman R, Lantinga M, Sangam S, et al.
Epi Info™, a Database and Statistics Program for Public Health Professionals. Atlanta, GA, USA: CDC; 2011.
Love J, Selker R, Marsman M, Jamil T, Dropmann D, Verhagen J, et al.
JASP: Graphical statistical software for common statistical designs. J Stat Softw 2019;88:1-7.
Sampasa-Kanyinga H, Zamorski MA, Colman I. The psychometric properties of the 10-item Kessler Psychological Distress Scale (K10) in Canadian military personnel. PLoS One 2018;13:e0196562.
Tavakol M, Dennick R. Making sense of Cronbach's alpha. Int J Med Educ 2011;2:53-5.
Furukawa TA, Kessler RC, Slade T, Andrews G. The performance of the K6 and K10 screening scales for psychological distress in the Australian National Survey of Mental Health and Well-Being. Psychol Med 2003;33:357-62.
Brooks RT, Beard J, Steel Z. Factor structure and interpretation of the K10. Psychol Assess 2006;18:62-70.
Li XM, Xiao WH, Yang P, Zhao HX. Psychological distress and cancer pain: Results from a controlled cross-sectional survey in China. Sci Rep 2017;7:39397.
Joshy G, Thandrayen J, Koczwara B, Butow P, Laidsaar-Powell R, Rankin N, et al.
Disability, psychological distress and quality of life in relation to cancer diagnosis and cancer type: Population-based Australian study of 22,505 cancer survivors and 244,000 people without cancer. BMC Med 2020;18:372.
Priede A, Hoyuela F, Umaran-Alfageme O, González-Blanch C. Cognitive factors related to distress in patients recently diagnosed with cancer. Psychooncology 2019;28:1987-94.
Donker T, Comijs H, Cuijpers P, Terluin B, Nolen W, Zitman F, et al.
The validity of the Dutch K10 and extended K10 screening scales for depressive and anxiety disorders. Psychiatry Res 2010;176:45-50.
Smith HR. Depression in cancer patients: Pathogenesis, implications and treatment (Review). Oncol Lett 2015;9:1509-14.
Hasler G. Pathophysiology of depression: Do we have any solid evidence of interest to clinicians? World Psychiatry 2010;9:155-61.
Ströhle A, Gensichen J, Domschke K. The diagnosis and treatment of anxiety disorders. Dtsch Arztebl Int 2018;155:611-20.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3]