Coping Modes With HIV Disease Predict Loss From HIV Study Cohort

Summary: The objective of this longitudinal study of 41 Swedish men with HIV disease was to assess what variables or combination of variables might predict loss to retention at 12 months. The variables of focus were of patients’ characteristics, quality of life, psychological, and medical characters. It appears that it is psychological variables relating to coping with HIV disease that are the best predictors of failure to retain respondents in this cohort. The responses to two items: ‘I have been doing things that I believe will improve my health (e.g. changed my diet)’ and ‘I feel like giving up’ correctly classified over 85% of cases lost to the study. These findings are worthy of replication and may prove useful in maximizing retention rates in longitudinal studies of the course of HIV disease and medication adherence and outcome.

Keywords: HIV infection, attrition, coping

Introduction

As increasing numbers of cohort studies are conducted with regard to the course of HIV disease and related medications, dropout (failure to retain in a study) from longitudinal studies poses a threat to the validity of the findings. To date, dropout has largely been described in demographic terms – the characteristics of individuals that may be associated with later dropout in a longitudinal study.1 Cozzi et al.1 emphasize the importance of understanding the biases of dropout to calculate attrition bias. We carried out a longitudinal study of 41 men with HIV disease, who were attending a hospital outpatient treatment clinic in Gteborg, Sweden, to assess what variables or combination of variables might predict loss to retention at 12 months. The variables of focus were of patients’ characteristics, quality of life (QoL), psychological, and medical character.

Methods

Patients and procedure

In connection to their regular hospital visit, a total of 69 asymptomatic and symptomatic patients with HIV infection at an outpatient HIV clinic in Gteborg, Sweden, were asked by HIV nurses for study participation. All patients had been diagnosed with HIV infection for at least 12 months. Patients who agreed to participate were fully informed about the study and signed an informed consent form. Fifty-eight patients (84%) agreed to participate.

Once informed consent was obtained, patients were given a self- administered questionnaire to be completed in the clinic. The questionnaire was sealed in an envelope and returned to the lead investigator to ensure anonymity and confidentiality. At 12 months follow-up visit, patients received an identical (apart from patients’ characteristics) questionnaire as the baseline one. All participants received regular clinic HIV treatment and counselling as appropriate in the course of the study.

The hospital committee for the protection of human subjects reviewed and approved study procedures and research instruments.

Measures

Patients’ characteristics Patients’ characteristics included age, education, employment, partner status, and years of HIV diagnosis.

Quality of life (QoL) QoL is in the study considered as a global and subjective condition that reflects the level of life satisfaction.2 QoL was measured by the item ‘My life as a whole is…’ from The Scale of Life Satisfaction.3 The response is on a six-point satisfaction Likert scale from ’1 = very dissatisfying’ to ’6 = very satisfying’. The Scale of Life Satisfaction also measures satisfaction with 10 QoL-related life domains (vocation, economy, leisure, social network, sexual life, family life, partner relationship, ability to manage self-care, health, and psychological functioning) on a similar scored scale.

Psychological distress, coping modes, and medical status Psychological distress refers here to hopelessness, a core characteristic of depression.4 It was assessed by the Beck Hopelessness Scale.4 The items are rated as True or False; the higher the score the higher degree of hopelessness. The internal reliability of the scale with the present sample was α = 0.91.

Coping modes were assessed by the Mental Adjustment to HIV Scale (MAHIV Scale).5 It is a modified version of the well-established Mental Adjustment to Cancer Scale.6 The scale items are rated on a four-point Likert-type response scale (α = ‘strongly disagree’ to 3 = ‘strongly agree’). In this study, three of the five subscales were being used, Helplessness/Hopelessness (α = 0.77), Belief in influencing the course of the HIV disease (α = 0.68), and Resiliency (α = 0.66). The Fatalism and Denial/Avoidance subscales were excluded due to low internal reliability. The first subscale, Helplessness/Hopelessness, is characterized by feelings of being at a loss as to what to do and out of control of the disease process. The second subscale, Belief in influencing the course of the HIV disease, is characterized by a belief in being able to influence the course of the disease, like exercise and diet and a positive state of mind. The third subscale, Resiliency, reflects people having a positive attitude and trying to fight the disease and not let it affect their lives.

Medical status, in terms of CD4 cell count, HIV disease phase, and HIV medication, was collected from the patient’s medical record. HIV disease phase incorporated two phases, the ‘asymptomatic’ and the ‘symptomatic’ disease phases. The latter was a combination of the ‘symptomatic’ and the ‘AIDS’ disease stages.

Statistical analysis

A total of 43 patients (74% of the enrolled patients) completed and returned the baseline questionnaire. The two female patients were dropped from the analysis. The baseline analysis was thus based on 41 men, of whom six self-identified as heterosexual, 30 as homosexual, and five as bisexual. At the 12 months follow-up, 24 (59%) of the male sample completed the questionnaire.

We compared patients’ characteristics, QoL, psychological, and medical characteristics for sexual orientation, and for drop-out patients with contactable patients using χ^sup 2^ (with Yates correction for discontinuity where appropriate) or t-tests for continuous data, all significance levels P ≤.05.

Stepwise linear discriminant analysis was carried out on the items of Helplessness/Hopelessness and Belief in influencing the course of the HIV disease MAHIV subscales, CD4 cells, and medication, given their significance in the bivariate analysis. Wilks’ λ was used to determine the significance level for entry into the equation. Pearson productmoment correlations were carried out between subscales of the MAHIV Scale to determine their degree of co-linearity. Analysis was conducted using SPSS version 11.

Results

At baseline, the mean age of the male white patients was 40.05 years, SD 10.55. Fourteen were single, the remainder was living with a partner; 21 had a high-school education, 17 had a university or other tertiary education. Seventeen were unemployed, 22 working, and two studying. All patients were infected with the virus through sexual transmission. The mean duration of being HIV diagnosed was six years, ranging from one to 14 years. Less than one-third (29%) of the patients were on no HIV medication, 42% on non-protease inhibitor (PI)-based combination therapy, and 29% on PI-based combination therapy. The Helplessness/Hopelessness coping mode scale was the only variable on which there was a sexual orientation significant difference; the heterosexual men scored higher than the gay men (t-test: 2.89, P = 0.006).

Twenty-four of the 41 men (59%) were contactable and willing to complete the follow-up questionnaire after 12 months: the remainder was considered lost to follow-up. Patients who dropped out had lower CD4 cells at baseline than those staying in the study (t-test: – 2.32, P = 0.03) and also more likely to be on PI-based combination therapy (t-test: 2.12, P = 0.04). While there was no significant association between the Resiliency subscale of the MAHIV scale and loss to retention, the Helplessness/Hopelessness and the Belief in influencing the course of the HIV disease subscales correlated significantly (t-test 2.30, P = 0.03 and t-test 1.92, P = 0.05, respectively). Pearson’s correlation between the Helplessness- Hopelessness subscale and Belief in influencing the course of the HIV disease subscale in this sample was -0.10, P = 0.53. No significant differences were found between the 17 dropouts and the 24 contactable patients for patients’ characteristics (age, sexual orientation, education, employment, partner status, years of being HIV diagnosed), QoL, psychological distress, and HIV disease phase.

Table 1 MAHIV Scale Items predicting drop-out respondents after 12 months

In the stepwise linear discriminant analysis, two items, ‘I have been doing things that I believe will improve my health (e.g. changed my diet)’ in the direction of ‘strongly agree’) (from the MAHIV subscale ‘Belief in influencing the course of the HIV disease’) and ‘I feel like I want to give up’ (in the same direction, from the MAHIV Helplessness/ Hopelessness subscale) significantly discriminated membership of the dropout versus contactable groups, correctly classifying over 85% of retention outcome (Table 1). Despite heterosexual male patients scoring significantly higher on the Helplessness/Hopelessness MAHIV subscale, there was no sexual orientation significant difference on \the item ‘I feel like I want to give up’.

Discussion

These data are based on a relatively small, male sample and thus caution should be exercised in generalizing beyond it. Nevertheless, there are two findings in these data that are worthy of replication and that may prove useful in maximizing retention rates in longitudinal studies of the course of HIV disease and medication adherence and outcome.

First, it appears that it is psychological variables relating to coping modes with HIV disease that are the best predictors of failure to retain respondents in this cohort. The responses to two items: Ι have been doing things that I believe will improve my health (e.g. changed my diet)’ (in the direction of ‘strongly agree’), and Ι feel like giving up’ (in the same direction), correctly classified over 85% of cases lost to the study. It may be worthwhile including these items in longitudinal studies to identify respondents at risk of loss to the study in order that additional attention and contact may be given to them, such as more frequent contact and development of a relationship with the study staff.

Second, these data also suggest that it is psychosocial variables related to the individual’s coping mode with the HIV disease that are more helpful predictors of retention than patients’ characteristics, psychological distress, QoL, or medical status. Further research in this area is warranted, particularly given recent data that suggest that psychosocial variables may also be useful predictors of adherence to medication regimens.7,8 Ability to identify those at risk for loss to the study may help reduce the threat to study validity and power resulting from loss of respondents from research cohorts.

Acknowledgements: The authors thank the women and men who participated in this study for their contribution, the nurses and psychologist Karl-Erik Lagberg at the HIV clinic, Sahlgrenska University Hospital/Sahlgrenska, Gteborg, Sweden for administering the questionnaires, and Lisbeth Hedelin, University of Gteborg, for statistical assistance. The study was funded by a grant from the Gteborg City Council.

References

1 Cozzi LA, Smith GD, Mocroft A, Sabin CA, Morris RW, Phillips AN. A practical approach to adjusting for attrition bias in HIV clinical trials with serial marker responses. AIDS 1998;12:1155-61

2 Day H, Jankey SG. Lessons from the literature toward a holistic model of quality of life In: Renwick R, Brown I, Nagler M, eds. Quality of Life in Health Promotion and Rehabilitation. Conceptual Approaches, Issues, and Applications. London: SAGE Publications, 1996;39-50

3 Fugl-Meyer AR, Brannholm IB, Fugl-Meyer KS. Happiness and domain specific satisfaction in adult northern Swedes. Clin Rehabil 1991;5:25-33

4 Beck A, Weissman A, Lester D, Trezler L. The measurement of pessimism: the Hopelessness scale. J Consult Clin Psychol 1974;42:861-5

5 Ross MW, Hunter C, Condon J. The mental adjustment HIV scale measurement and dimensions of response to AIDS/ HIV disease. AIDS Care 1994;6:407-11

6 Watson M, Greer S, Bliss JM. Development of a questionnaire measure of adjustment to cancer: the MAC scale. Psychol Med 1988;18:203-9

7 Nilsson Schonnesson L, Ross MW, Williams M. The HIVmedication self-reported non-adherence reasons (SNAR) index and its underlying psychological dimensions. AIDS Behav 2004;8:293-301

8 Ammassari A, Trotta MP, Murri R. Correlates and predictors of adherence to highly active antiretroviral therapy: overview of published literature. J Acquir Immune Defic Syndr 2002;31:S123-7

(Accepted 28 May 2004)

Lena Nilsson Schnnesson PhD1, Michael W Ross PhD MPH2 and Ing- Marie Bergbrant MD3

1 Gay Men’s Health Clinic, Gay Men’s Health Center, Sder Hospital/ Karolinska Institute, Sdersjukhuset, SE-118 83 Stockholm, Sweden; 2 WHO Center for Health Promotion and Prevention Research, School of Public Health, University of Texas, Houston, TX, USA; 3 Department of Dermatology, Sahlgrenska University Hospital/Sahlgrenska, Goteborg, Sweden

Correspondence to: Dr Lena Nilsson Schnnesson

Email: lena.nilsson-schonnesson@karolinska.se

Copyright Royal Society of Medicine Press Ltd. Jul 2005