Article Text
PDF
Abstract
Objectives There is a paucity of data reporting the prevalence and burden of pain and symptoms among HIV patients in the era of antiretroviral therapy (ART). This study aimed to measure symptom prevalence and determine associations with key variables: demographics, treatment status, adherence and risk behaviours.
Design Cross-sectional self-completion questionnaire in five HIV outpatient clinics in London and the south east UK.
Methods Consecutive patients were invited to participate, responding to clinical and behavioural variables including the memorial symptom assessment schedule (short form). Four multivariable models examined the relationship between dependent variables of psychological, physical, global symptom burden scores, the number of symptoms and key independent variables. 778 patients participated; response rate 77% of all patients approached.
Results Both physical and psychological symptoms were highly prevalent: in the previous 7 days, 70.8% experienced lack of energy, 69.9% worry, 53.6% diarrhoea, 53.5% sexual dysfunction and 53.2% pain. In multivariable analysis, both unprotected sexual intercourse with a partner of unknown status, and poorer ART treatment adherence were significantly and independently associated with psychological symptom burden. Lower educational achievement was significantly associated with increasing physical, psychological and global symptom burden and with higher number of symptoms. Being on ART was not associated with any symptom distress measure.
Conclusions In the era of treatment, patients continue to experience high prevalence and burden of psychological and physical symptoms, which are not associated with treatment status. Attention to these distressing problems is essential and may enhance quality of life and adherence and minimise risk behaviour. Symptoms are highly prevalent among HIV outpatients and associated with poor adherence and unprotected sexual intercourse.
- Adherence
- HIV
- pain
- risk
- symptoms
Statistics from Altmetric.com
HIV disease is associated with highly prevalent and distressing symptoms from the point of diagnosis.1 A significant proportion of patients receiving antiretroviral therapy (ART) experience drug toxicities such as peripheral neuropathy and gastrointestinal problems,2 requiring drug discontinuation or treatment change within and between classes of ART.3 However, the vast majority of data on pain and symptom prevalence has been reported from the pre-ART era on patients with AIDS-defining illnesses, and it has been difficult to quantify the relative contribution of treatment to the presence and severity of symptoms.4
The WHO advocates pain and symptom control as an essential component of HIV clinical care.5 However, the evidence suggests inadequate clinical attention to pain and symptom assessment and management, as clinicians detect threefold fewer HIV-related symptoms than their patients report,6 and patients perceive that their symptoms remain untreated.7 8
In order to maximise the clinical benefit from ART, adherence must be optimised and movement between classes of drugs minimised so that future treatment options can be retained. In order to achieve this, ART-related symptomatology must be taken into account. Furthermore, clinicians need to be able to identify prevalent symptoms, and which patients may experience symptom distress, as well as the non-clinical potential ramifications of symptom distress such as risk behaviour and adherence variation. Not only is there currently a lack of focus on pain and symptom prevalence in the era of ART, but these variables have not been assessed in relation to the key clinical and behavioural components of care such as treatment adherence and behavioural risk reduction. Investigation into the prevalence and correlates of symptoms is required to inform clinical practice and thereby improve clinical outcomes and quality of life.1 9 10
Methods
Design and recruitment
This study aimed to measure the prevalence and distress related to physical and psychological symptoms, and to determine the relationship between these and demographic variables, ART use and adherence, risk behaviour and HIV status disclosure.
This cross-sectional questionnaire study was conducted in five outpatient clinics in London and the south east of England during 2005–6. Consecutive attending patients were approached and invited to participate (n=1007). The inclusion criteria (HIV-positive adults with sufficient English language and cognitive/physical ability to self-complete) were fulfilled by 904 patients (90%) who received a questionnaire, of which 778 were completed and returned (77% of all patients, 86% of those eligible to receive a questionnaire). Multicentre UK ethical approval was granted by the Royal Free Hospital Clinical Ethics Committee ref 05/Q1907/26.
Data collection
The questionnaire consisted of four core components of items that were selected to meet the aims of the study in line with the existing literature.
First, a demographic component (addressing age, gender, ethnicity, employment, country of birth, sexuality and education).
Second, a behavioural component (unprotected sexual intercourse with a partner of unknown status in the previous 3 months; treatment optimism and infectiousness optimism on a standardised scale11; and disclosure of HIV status to others based on Kalichman et al).12
Third, treatment items (current treatment status was recorded as naive, on first regimen, switched regimen or stopped; treatment adherence based on a 7-day self-report recall on three levels: fully adherent, partly adherent or non-adherent, a composite measure according to whether all doses were taken, at the correct time and in line with dietary requirements).13
Fourth, symptoms were measured using the memorial symptom assessment scale (MSAS; short form), a standardised scale reporting on symptom prevalence in the previous 7 days, with three subscale indices of physical symptom distress, psychological symptom distress and global distress index.14 Each of these three subscales has a possible score range of 0–4. This well-validated multidimensional instrument captures the presence and distress of 26 physical and six psychological symptoms. The total number of symptoms (out of a possible 32) was also ascertained. Although it was originally validated among cancer patients, it is commonly applied in a range of progressive conditions, and has the advantage of enabling comparison between patient populations.
Analysis
Descriptive statistics were used to report the entire sample characteristics.
The following analysis was then conducted on the full sample.
Univariate linear regression analyses were conducted to test the association of various factors with the following four dependent variables: (1) physical symptom distress subscale; (2) psychological symptom distress subscale; (3) global symptom distress subscale and (4) the number of symptoms. Each of these four dependent variables formed a separate model in the subsequent multivariable analyses described below.
Independent variables entered in univariate analysis for the whole sample were: age, gender/sexuality, ethnicity, UK born or not, education, unprotected intercourse, disclosure and current treatment status. For each model gender/sexuality, ethnicity, education and treatment were collapsed to two levels: gay male/heterosexual, university educated/not, Caucasian/not Caucasian and on ART treatment/not on treatment.
Subsequently, for the subsample of patients currently on ART treatment, the analysis was repeated for those patients only with additional treatment-related independent variables entered: treatment switching as two levels (switched/not switched) and adherence entered as a three-level variable: fully adherent/partly adherent/non-adherent. Current treatment status was not entered into these models.
Following each univariate regression, multivariable regression models were constructed. Independent variables from the univariate analyses above were entered stepwise into the multivariable model if significant at the 25% level.15 Cases with missing data were excluded from the multivariable models.
The multivariable analysis was conducted first on the entire sample, then for the subsample currently on ART.
Results
Sample characteristics
The sample mean age was 40.4 years (SD 8.6). The majority identified as gay or bisexual men (n=496, 65.7%), with n=183 (24.2%) women and n=76 (10.1%) heterosexual men (missing n=23). The majority were Caucasian (n=513, 67.3%), with n=188 black (24.7%), n=23 Asian (3.0%) and n=38 (5.0%) mixed/other (missing n=16). The majority were UK born (396, 50.9%). Educational attainment was reported as follows: no qualifications (n=67, 9.0%); exams attained at age 16 years n=190 (25.4%); exams attained at age 18 years n=154 (20.6%), university degree or above n=336 (45.0%), missing n=31.
Treatment variables
Most patients were currently on ART (524, 67.4%), with 155 (19.9%) treatment naive (the remaining 99 having stopped treatment). Adherence in the preceding 7 days was as follows: fully adherent 41.5%; partly adherent 36.1%; non-adherent 22.4%. HIV status had not been disclosed to anyone by 47 respondents (6%).Unprotected sexual intercourse in the previous 3 months with a person who respondents were not sure was also HIV infected was reported by 85 of the sample (10.9%).
With respect to treatment switching, of those on ART, 161 (32.7%) had not switched from their first treatment regimen, 135 (27.4%) had switched once, 196 (39.8%) reported multiple switches (32 missing).
Symptomatology
The mean number of reported symptoms was 17.9 (SD 12.1), the mean (SD) symptom distress subscale indices were physical distress 0.81 (0.7), psychological distress 1.34 (1.0) and global distress 1.16 (0.8).
The prevalence and distress of the 10 most commonly reported symptoms are described in table 1. Manifestations of fatigue were most common, namely lack of energy (70.8%) and feeling drowsy/tired (67.5%). Problems related to psychological distress included difficulty sleeping (61.8%), difficulty concentrating (60.7%), feeling worried (69.9%), feeling sad (66.3%) and feeling irritable (56.6%). Other common physical symptoms were diarrhoea (53.6%), problems with sexual activity (53.5%) and pain (53.2%).
The 10 physical and psychological symptoms with highest 7-day prevalence and their associated distress (MSAS-SF) (n=778)
Multivariable analysis
For the full sample (n=778), table 2 shows univariable associations and table 3 shows multivariable regressions. In the multivariable models, lack of university-level education was significantly associated with poorer outcomes in all symptom distress indices; regression coefficients (b) for university level compared with other education were as follows: b=−0.111, p=0.004 for physical distress, b=−0.096, p=0.017 for psychological distress, b=−0.101, p=0.007 for global distress, b=−0.090, p=0.021 for total number of symptoms. Identifying as Caucasian was associated with lower global distress (b=−0.076, p=0.040). Having disclosed HIV status was significantly associated with a lower number of symptoms (b=−0.090, p=0.021). Reporting unprotected intercourse in the previous 3 months with a person not known to be also HIV positive was associated with greater psychological symptom distress (b=0.080, p=0.047). Interestingly, currently being on ART was not significantly associated with any of the symptom measures.
Associations of factors with symptom measures: univariable analysis whole sample n=778
Associations of factors with symptom measures: multivariable analysis models for whole sample n=778
Tables 4 and 5 show the analysis restricted to patients currently on ART (n=524). In multivariable models (table 5), poorer ART adherence was significantly associated with both psychological (b for trend across adherence categories 0.126, p=0.001) and global symptom distress (b=0.125, p=0.006). Having ever switched treatment was significantly associated with higher physical symptom distress (b=0.135, p=0.003), higher psychological symptom distress (b=0.126, p=0.006), higher global symptom distress (b=0.152, p=0.001) and a greater number of symptoms (b=0.132, p=0.013). Being UK born and university education were also independently associated with lower physical symptom burden.
Associations of factors with symptom measures: univariable analysis only for those on treatment n=524
Associations of factors with symptom measures: multivariable models only for those on treatment n=524
Discussion
The data, among outpatient HIV clinic attendees, reveal high 7-day prevalence and associated distress of burdensome symptoms. Psychological problems and pain were of particular concern. A symptom prevalence grid comparing evidence for end-stage patients reported on three of the most prevalent symptoms also reported in this study.16 Prevalence in our sample was at the top end of the reported range for both diarrhoea and fatigue, and in the mid-range for pain, compared with patients with advanced malignant, renal and pulmonary disease. Therefore, the symptom control needs of HIV outpatients, including those on treatment, are comparable with those with advanced malignant and non-malignant progressive diseases. This high prevalence of distressing symptoms should be taken into account to inform assessment and management within HIV clinical services.
The associations between symptom prevalence/burden and demographic, clinical and behavioural variables have identified significant relationships that should be taken account of to inform symptom assessment, adherence maximisation and risk reduction. The associations with being non-UK born/not Caucasian may be a reflection of the tendency for this population to be diagnosed with more advanced disease. It may also be that persons born outside the UK, where HIV stigma is more prevalent, are less likely to disclose their infection and therefore experience a high symptom burden. The association between having switched regimen and a greater number of symptoms may be due to those with more symptoms being more likely to wish to change treatment, having to take regimens with a higher side-effect burden, or simply having more advanced disease.
It is of particular note and importance that the data have shown that currently being on ART was not associated with having fewer symptoms or lower symptom distress, strongly indicating the continuing need for clinical attention to symptoms after initiating ART. Longitudinal studies are needed to identify causal relationships between initiating/maintaining specific treatment regimens and both physical and psychological symptomatology.
There are several potential limitations to our findings. First, although the response rate was high (77% of all clinic attendees, 86% of eligible attendees), a small number of patients did not meet the inclusion criteria (ie, ability to self-complete, in line with ethical review requirements). The period of recruitment (3 months) would have enabled most patients to be approached in outpatient settings. The cross-sectional nature of the study can identify associations, but cannot identify causal relationships. Although the response rate is high, we do not have data to compare the prevalence of symptoms between responders and non-responders. Additional data in future studies on the relationship between biological markers of disease (CD4 cell count, viral load) and symptomatology would add further understanding of the aetiology of these problems. We would also advocate further studies to investigate these relationships in resource-poor settings. Finally, as a result of the early provision of needle exchange in the UK, we have a very small population of people infected by intravenous drug use. Further study is warranted in countries where this population constitutes a greater proportion of HIV patients.
The body of evidence on the symptom control needs of people living with HIV disease in the ART era has been limited, and has not previously been investigated in relation to other clinical and behavioural variables. There are many barriers to effective symptom control for people with HIV disease, including the lack of awareness of symptoms associated with HIV disease, patient belief that pain should be expected following an HIV diagnosis, and a focus on purely virological outcomes.17 Greater clinical attention must be paid to the high prevalence of distressing psychological symptoms experienced by the majority of people living with HIV, as a systematic review of the evidence demonstrates that symptoms can be effectively managed.18
It is essential that quality management of HIV disease routinely assesses these distressing problems, so that key outcomes of risk behaviours and adherence may be optimally influenced. An essential step towards holistic management of the patient is to ensure that care protocols routinely ask about the full range of problems that the patient may experience. Physician-led identification of problems is likely to lead to the improved identification and management of problems than by expectations that patients will present with their physical and psychological symptoms. Effective treatment strategies should be implemented through integrated patient management between HIV clinicians, mental health providers and symptom management specialists. Generalist palliation of problems should be a core skill among all clinicians, with clear referral guidelines to, and knowledge of, local symptom control and palliation teams. In terms of management strategies for control of the highly prevalent range of psychological problems (and manifestations such as fatigue), systematic reviews of the evidence are urgently required to identify the most effective approaches. As HIV prognosis has greatly improved and a chronic disease model prevails, interventions that are appropriate to longer-term coping are required. It is clear that despite advances in universal treatment, the patient burden of disease remains high, and outcomes are unlikely to be improved without careful attention to the patient experience of disease and a clinical focus beyond virology.
Key messages
The burden of psychological and physical symptoms among HIV outpatients with treatment access is seldom researched yet is essential to improved management and patient outcomes.
Among 778 UK outpatients, the 7-day period symptom prevalence was high. Being on treatment was not associated with lower symptom distress. Psychological symptoms were associated with poorer adherence and sexual risk behaviour.
Clinical care should routinely assess multidimensional symptoms as these are controllable and may improve outcomes such as treatment response, transmission and quality of life.
Acknowledgments
The authors would like to acknowledge the contribution of Gilly Arthur and Sarah Zetler, also research assistance from Amanda Jayakody, research nurses at participating clinics, and all survey respondents. They are grateful to Lucy Bradley for manuscript management.
References
Footnotes
Funding This research was assisted with an unrestricted educational grant from GlaxoSmithKline, with input from the Adherence Strategy Group.
Competing interests None.
Ethics approval This study was conducted with the approval of the Royal Free Hospital University College Medical School Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.