Identification of over-compliant patients and pill dumpers with the mixed pill count method at an antiretroviral treatment clinic

African Journal for Physical, Health Education, Recreation and Dance (AJPHERD) Supplement 1 (September), 2014, pp. 105-115. Identification of over-co...
Author: Clarissa Kelly
2 downloads 3 Views 102KB Size
African Journal for Physical, Health Education, Recreation and Dance (AJPHERD) Supplement 1 (September), 2014, pp. 105-115.

Identification of over-compliant patients and pill dumpers with the ‘mixed’ pill count method at an antiretroviral treatment clinic TITILOPE A. AWOLOLA1 JOHANNA SUMMERS1 AND SUZANNE JOHNSON2

C.

MEYER1,

BEVERLEY

1

Department of Pharmacy, University of Limpopo (Medunsa Campus), South Africa, P.O. Box 218, Medunsa 0204, South Africa. E-mail: [email protected] 2 Foundation for Professional Development, Struland Office Park, The Willows, Pretoria 0184, South Africa

Abstract Near perfect levels of adherence to antiretroviral treatment (ART) are required for maximal viral suppression and avoidance of drug resistance. Pill counts are an objective method of measuring adherence, but can be invalidated if patients manipulate the number of tablets returned. This research aimed to investigate whether the ‘mixed’ pill count method was capable of identifying patients who were engaging in pill manipulation (dumping) at a public sector antiretroviral (ARV) clinic in South Africa. At baseline, 370 adults on ART were recruited. On the first return visit, a standard/normal pill count was used to calculate adherence and extra tablets were dispensed without the patients’ knowledge. At the second return visit, adherence was calculated with the ‘mixed’ pill count. Patients were grouped into three categories based on calculated adherence: truthfully non-adherent, adherent and over-compliant. Three hundred and forty-four (92.9%) patients completed the study. At the second return visit, with the ‘mixed’ pill count, 43 patients (12.5%) were identified as over-compliant based on the adherence calculated for their regimens. The ‘mixed’ pill count identified pill dumpers, although overall adherence was better on the second return visit compared to the first. The significance of this research lies in the fact that it indicates that continuous individual patient counselling and informing patients of their exact adherence score contributed to motivating them to adhere well to their treatment. Keywords: Dump, mask, over-compliance, adherence, antiretroviral treatment, mixed pill count. How to cite this article: Awolola, T.A., Meyer, J.C., Summers, B. & Johnson, S. (2014). Identification of over-compliant patients and pill dumpers with the ‘mixed’ pill count method at an antiretroviral treatment clinic. African Journal for Physical, Health Education, Recreation and Dance, September (Supplement 1), 105-115.

Introduction Strict adherence to antiretroviral treatment (ART) is essential for treatment success (Ford et al., 2010). Many patients on ART struggle to take their medicines consistently for various reasons. It is therefore crucial that health care providers use efficient and reliable tools to identify sub-optimal adherence and address the problem timeously.

106 Awolola, Meyer, Summers and Johnson Current standard measures to assess adherence to antiretrovirals (ARVs) such as pill counts, directly administered antiretroviral therapy (DAART), therapeutic drug monitoring, self-reported adherence and electronic medication monitoring systems (MEMS), all have advantages and limitations (Nachega, Knowlton, Deluca & Schoeman, 2006; Deschamps et al., 2008; Altice & Springer, 2010). Presently no accepted ‘gold standard’ exists by which to measure adherence to ART and the above-mentioned methods are used individually or in combination (Berg & Arnsten, 2006; Chesney, 2006; Ross-Degnan et al., 2010; Steel, Nwokike, Joshi & Ntengu, 2011). The need for efficient and reliable measures of adherence has initiated debate on how best to measure medication-taking behaviour in public sector ART settings. Although the pill count is an indirect, objective method to assess adherence, it tends to overestimate adherence and does not indicate true medication-taking behaviour in terms of timing and dosage (Ross-Degnan et al., 2010). Evidence from the literature (Steel et al., 2011) and anecdotal evidence from some ART sites in South Africa, suggest that the pill count method can be invalidated by patients if they manipulate the number of tablets returned to the clinic, in an attempt to appear adherent at the follow-up visit. Allegedly, these patients deliberately fail to return all their tablets or discard (dump) the remaining tablets before coming to the clinic for their monthly pill count and medication refill. To detect the conjectured manipulations, the ‘mixed’ pill count method was implemented to measure adherence in this study. The ‘mixed’ pill count is a practice whereby the pharmacist deliberately dispenses extra tablets to a patient (a predetermined ‘n’ number of extra tablets more than the usual 30-day supply) and records the number of extra tablets in the patient’s medication file without the patient’s knowledge. When the patient returns for a repeat prescription at the follow-up visit, the tablets remaining in the container are counted and the ‘n’ number of extra tablets dispensed is taken into consideration. The hypothesised advantage of the ‘mixed’ pill count over the ‘normal’ pill count lies in the fact that patients are not informed that extra tablets have been dispensed. As a result, it is assumed that the following three scenarios could ensue: • An adherent patient would return to the clinic with the extra tablets in the container because he/she is sure that all the tablets were taken as prescribed, resulting in an adherence rating of 100%. • A truthfully non-adherent patient would return to the clinic with the extra tablets (and more) in the container because he/she is not attempting to mask nonadherence and, therefore, would have no need to discard any tablets (including the extra ones), resulting in an adherence rating of 100%. In all these instances, the extra tablets should act to safeguard the validity of the pill count. Inability to account for the extra tablets would allow the pharmacist to easily identify patients who may have attempted to mask non-adherence by dumping tablets and also allow the health care worker to probe for and address the issues behind non-adherence masking. The end result would be improved ability to identify non- adherence, which previously would have been successfully masked. This study aimed to investigate if the ‘mixed’ pill count was capable of identifying patients who were engaging in pill manipulation. Methodology Study design A prospective and longitudinal design was followed over a period of six months at Ntshembo public sector ARV Clinic in Mamelodi Hospital, Gauteng Province, South Africa. Study population and sampling All treatment-experienced patients were invited to participate in the study. Patients were informed that adherence would be monitored with a pill count, but were semi-blinded, as they did not know that extra tablets would be added to their containers during the second study visit (i.e. first return visit). Purposive sampling was used to select participants who were HIV-positive adults ≥18 years, on first-line ARV regimens and patients who would be receiving their ARVs at the study site for the full duration of the study period. Patients who were on TB treatment, pregnant patients and those who were down-referred to another ART site were excluded from this study. At the end of the recruitment period, 370 patients were enrolled into the study. Data collection procedures and instruments At enrolment (baseline visit) participants were given a normal pill count, all tablets returned were collected from them and each patient was dispensed the usual supply of his/her ARV medication. The correct number of tablets consumed would then be used for adherence calculation at the two return visits. Patients were reminded to bring their medication containers during return visits. A questionnaire, available in English and Setswana, was used to record demographic and clinical data.

108 Awolola, Meyer, Summers and Johnson On Return Visit 1, patients received a normal pill count and extra dosage units of each ARV medication were dispensed to them without their knowledge. At Return Visit 2, patients received a ‘mixed’ pill count. A pill count sheet was used for the normal pill count as well as the ‘mixed’ pill count, to record the number of tablets that the patients returned to the clinic. Pill count results were calculated as a percentage of the expected consumption as follows: % Adherence = Number of tablets taken home – number of tablets returned X 100 Number of tablets that should have been taken

The pill count sheet designed specifically for this study was used during the study period in addition to the in-house pill count sheet normally used by the clinic. At Return Visits 1 and 2 patients were informed of their exact adherence percentage and counselled by the clinic counsellors according to normal clinic procedures. Pilot study This study was pilot-tested amongst 29 patients at Masibambane ARV Clinic, Tshwane District Hospital, an ART site different from the main study site (Adeyinka, Meyer, Summers & Johnson, 2010). According to the pilot study results, the ‘mixed’ pill count method appeared to be capable of detecting overcompliance and pill dumpers. The concept was therefore used as part of normal practice during the main study. Statistical analysis Data were analysed with Statistical Analysis System (SAS®) software. The Fisher’s exact- and McNemar’s tests were performed to determine any significant difference in adherence between the variables compared, with p≤0.05 regarded as statistically significant. Demographic data were summarised and expressed as means and percentages. Ethical considerations The University of Limpopo, Medunsa Campus Research and Ethics Committee granted ethical clearance for the study. Approval to conduct the pilot study and main study was granted by the Chief Executive Officer of the respective study sites. All participants provided written consent prior to enrolment and were assured of confidentiality of data.

Identification of over-compliant patients and pill dumpers 109 Results Three-hundred-and-forty-four patients (93.0%; n=370) completed the study. Dropout (7%) included 11 patients who switched regimen, eight patients who defaulted during the course of the study, and seven patients who were transferred to another ART clinic. Socio-demographic characteristics Socio-demographic characteristics at baseline are summarised in Table 1. The mean age of the study population (n=344) was 40.1±9.11 years, with females predominating (205; 59.6%). On average, participants were on ART for 30 months with a median time of 24 months. Table 1: Socio-demographic characteristics of the study population at baseline Characteristics Male (n=139) Female (n=205) No (%)1 No (%)1 18-30 12 (8.6%) 39 (19.0%) Age group (years) 31-40 55 (39.6%) 90 (43.9%) 41-50 49 (35.3%) 48 (23.4%) >50 23 (16.5%) 28 (13.7%) Black 135 (97.1%) 202 (98.5%) Race Coloured 4 (2.9%) 3 (1.5%) Afrikaans 5 (3.6%) 5 (2.4%) Language Zulu 25 (18.0%) 32 (15.6%) Tswana 10 (7.2%) 17 (8.3%) Sotho 33 (23.7%) 57 (27.8%) 66 (47.5%) 94 (45.9%) Other2 None / primary not 13 (9.4%) 22 (10.7%) Level of completed education Primary completed 69 (49.6%) 102 (49.8%) Secondary completed 49 (35.3%) 72 (35.1%) Tertiary / vocational 8 (5.8%) 9 (4.4%) 77 (55.4%) 82 (40%) Employment Employed status Unemployed 62 (44.6%) 123 (60%) Single 83 (59.7%) 135 (65.9%) Marital status Married 38 (27.3%) 47 (22.9%) Widowed 8 (5.8%) 16 (7.8%) Divorced 10 (7.2%) 7 (3.4%) 100% adherence) per individual ARV for the two pill count methods (n=43)

ARV medication Stavudine Lamivudine Efavirenz Nevirapine Fixed-dose combination# Tenofovir

Normal pill count at Return Visit 1 n* No (%) 28 7 (25.0%) 28 7 (25.0%) 34 5 (14.7%) 10 5 (50.0%) 12 2 (16.7%) 4 2 (50.0%)

Mixed pill count at Return Visit 2 n No (%) 24 21 (87.5%) 27 24 (88.9%) 32 21 (65.6%) 12 9 (75.0%) 15 14 (93.3%) 5 2 (40.0%)

P-value (Fisher’s exact test)

Suggest Documents