|
nif:isString
|
-
This pre-post control group study design [10] was administered as a: a) retrospective clinical records study and; b) prospective controlled study. The study covered two administrative divisions of Nairobi East District. Randomization Software EASY RA1 Easy Randomizer Version 4.1, State University of Michigan, USA was used and Njiru was randomly designated as the intervention division and Makadara as the control area. According to health records, these two divisions were comparable in terms of population numbers, staff numbers, personal and background characteristics of the staff and workloads at the health facilities. Each division comprised four health facilities. The study included all PHWs (n = 32 in intervention and n = 25 in control division) and all CHWs (n = 411 and 404 in the intervention and control division, respectively). Flow of the participants, as well as activities and types of data that were collected at various stages of the study, are described below and illustrated in Boxes S1 and S2.
Improvement of Knowledge in Recognition of HROLs by Health Care Providers: Radboud University Nijmegen Medical Center, Department of Global Oral Health and the University of Nairobi Dental School collaboratively developed two one-day training programs, to equip the health workers with basic knowledge and skills for recognition and management of (HIV-related) oral lesions, as well as reminding programs, as described elsewhere [8], [9]. PowerPoint presentations contained simplified text and selected photographs of common (HIV-related) oral lesions (pseudo membranous candidiasis, erythematous candidiasis, angular cheilitis, oral hairy leukoplakia, necrotizing ulcerative gingivitis, Kaposi’s sarcoma, parotid enlargement, herpes zoster, dental caries, periodontitis, dental fluorosis) [1], [3], [11] to enhance healthcare provider ability to recognize these lesions. During the training, PHWs also received practical training, involving HIV-infected patients with HROLs, to improve their clinical skills. In the first three months after the training, PHWs were visited and reminded to review patients with questionable lesions and use PowerPoint presentations containing photographs of HROLs to update their knowledge. Barrier factors affecting their performance were also assessed. All health centers received monthly telephone reminders about the correct use of routine recording tools. CHWs received an additional module on communication skills and use of educational brochures and posters, to enable them to do community mobilization.
A baseline written assessment was done, to analyse background characteristics and baseline knowledge amongst PHWs and CHWs in the intervention and control divisions, as described elsewhere [5]. On the one hand, oral health care tasks are relatively new among professional primary health care providers as this topic is neither emphasised during pre-service nor in-service training. On the other hand, low literacy levels in this community and lack of formal medical training among CHWs explains the numerous traditional beliefs and practices as well as misconceptions regarding HIV related (oral) diseases. To gather as much information as possible from both health workers and the CHWs, open ended questions were additionally used. The PHW questionnaire covered knowledge of symptoms of HROLs, the clinical appearance of HIV-suspected conditions, general dental knowledge, knowledge of OPC, common appearances of OPC, knowledge of periodontitis, causes of dental caries, past training in oral health topics and clinical experiences (Table 1). The 5 domains in the CHW questionnaire were: general oral health knowledge, knowledge of HROLs, opinions regarding oral health problems, encounter of HROLs in the community and current care at community level (Table 2). Both questionnaires were in English. The CHW questionnaire had an additional translated version, in their local language, to enable them to express their views accurately. General results of this assessment have been presented elsewhere [5], [6]. These questionnaires were again administered; 9 months after the training for PHWs and 6 months after training for CHWs (post-training assessment). Training and evaluation of PHWs was performed between February 2010 and November 2010. Thereafter, CHW training was done and evaluated (February 2011 to September 2011).
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0090927.t001 Pre- and post-training single item questions and mean domain scores, standard deviations (sd), p-values and maximum scores in the PHW questionnaire for intervention and control division. 1Intervention group (I): n = 32; Control group(C): n = 25.2Intervention group (I): n = 25; Control group(C): n = 15.HROLs – HIV related oral lesions. *domains.
Table data removed from full text. Table identifier and caption: 10.1371/journal.pone.0090927.t002 Pre- and post-training mean domain scores, standard deviations (sd), maximum scores and p-values in the CHW questionnaire for intervention and control divisions. 1intervention group (I): n = 32; control group(C): n = 27.2intervention group (I): n = 27; control group(C): n = 15.HROL – HIV related orofacial lesions.
Performance of oral Examinations and Identification of HROLs by PHWs: PHWs routinely record outpatient consultation details, including oral pathologies (dental decay, periodontal diseases, malocclusion and (HIV-related) oral mucosal lesions) in daily tally sheets and outpatient registers. These indicators were retrospectively and manually extracted from the clinical records from the year prior to the training as well as after training, in the intervention and control divisions. Clinical records were studied in assessing whether or not an oral examination had been performed. This was subject to limitations such as: failure of PHWs to record oral examinations if they did not find relevant oral pathology; some orofacial lesions such herpes zoster had been diagnosed without oral examination. Therefore, additional assessments were done by structured exit interviews (n = 924) in the intervention area and (n = 666) control division, a month before the training and thereafter on patients leaving the health facilities, to assess whether an oral examination had been performed by PHWs (Box S1).
Referral by CHWs of Patients from the Community with HIV-related Oral Lesions: At recruitment, all CHW and PHWs had been introduced to their roles: educating the community, referral of patients to PHWs for further care and use of the government- owned CHW service delivery log book (CHWSDLB) for recording all referrals that they made from the community. The CHWSDLB was modified, through addition of two columns for the capture of data on the number of referred HROLs. CHWs received short messages on their mobile phones, known to be effective reminders in this setting [12], and supervisory visits to remind them to record and submit monthly data from the community through community health extension workers. Data were collected two months before the aforementioned CHW training program and throughout the succeeding 6 months after the training.
HIV Testing Rates and Capacity to Identify High Risk Patients: The expectation was that: a) trained PHWs would refer more patients with HROLs from outpatient consultation rooms to the laboratory for HIV testing and; b) trained CHWs would refer more patients with HROLs from the community to the PHWs and consequently further increase the number of patients with HROLs referred to the laboratory for HIV testing. This would increase overall HIV testing rates. PHWs refer patients to the laboratory within the same health facility for HIV testing. There a rapid HIV diagnostic test is done. All laboratories at each health facility keep records including the number and results of the tests. However, reasons for referral are routinely not included. The latter, including specific types of HROLs referred, were included after the training. Routine HIV laboratory data from voluntary counseling and testing (VCT) clinics, from prevention of mother to child transmission of HIV infection (PMTCT) and from diagnostic counseling and testing (DTC) in tuberculosis clinics were not part of the present analysis. In cases where patients with HROLs were referred by PHWs from these clinics to the laboratory for testing, the data was captured in the laboratory register.
Data were checked for completeness and consistency before entry into Excel files. Data analyses were conducted, using SAS software (version 9.2, SAS Institute, Cary, NC, USA) by JM (biostatistician at Radboud University Nijmegen Medical Centre). Background characteristics were described, using frequency and descriptive statistics. Chi-square tests were used in comparing differences in percentages. Pre-test means were analyzed, using Student-T-tests, whereas a multiple regression model with correction for pre-test values was used for analyzing post-test means. For all tests, a significance level of 0.05 was used. For effect assessment the data listed below were compared in both divisions, before and after the PHW and CHW trainings.
In both divisions, and for both test and control groups, student-T-Tests were used to compare baseline knowledge of lost-to-follow-up PHWs and CHWs with that of PHWs and CHWs who completed the post-training questionnaire.
Assesment of Knowledge in Recognition of HROLs by Health Care Providers: A panel of two experienced dentists and native speakers of both languages independently coded and iteratively harmonized open-ended data in the PHW and CHW post-study written assessments, as no prototype questionnaire was available and as much information was required to make a tailor-made responsive training programme. To generate an initial codebook for open-ended questions, the panel discussed to reach consensus on correct and wrong responses for open ended questionnaires with multiple responses. To improve the code book the scope of right and wrong answers was further defined using the questionnaire responses. The coders identified and discussed unsatisfactory responses, partially answered questionnaires, ambiguous statements as well as wordings of correct responses. Codes were assigned to agreed correct responses, wrong responses, unanswered questions and ‘I do not know’ responses. The number of unanswered questions was negligible since the respondents had the ‘I do not know’ option. In addition research assistants checked for completeness of the scripts during questionnaire administration. For analysis, unanswered questions and ‘I do not know’ questions were regarded as lack of knowledge and were therefore considered as wrong responses. The two open ended questions that had no right or wrong answer, as that they needed ‘further explanation’ were analysed separately. After coding the first seven scripts, results were checked for agreements and differences. Disagreements were discussed until consensus was reached; 90% agreement between the two coders permitted reliable coding of the remaining transcripts. Data were analysed in line with the aforementioned domains. After correction for baseline differences, pre- and post-training study mean scores for PHWs and CHWs were compared at (p = 0.05). Pre-test means were analyzed through use of Student-T-tests, whereas a multiple regression model with correction for pre-test values was used for analysis of post-test means.
Assessment of Performance of Oral Examination and Identification of HROLs by PHWs: Percentages were used to compare the difference in proportions of the number of outpatient consultations that included an oral examination, the number of diagnoses of HROLs and the number of outpatient consultations. Assessment of CHW Referrals of Patients with HIV-related Oral Lesions from the Community: Numbers of HROLs referred from the community to the health facilities were compared, as well as the numbers of HROLs diagnosed by PHWs in the outpatient clinics. Assessment of HIV Testing Rates and Capacity to Identify High Risk Patients: Proportions of a) outpatient consultations that included an HIV test b) HROLs diagnosed in the outpatient clinics that received an HIV test and c) HIV-positive test results in oral and non-oral lesions were expressed as percentages. Chi square tests were used in comparing differences in percentages at p<0.05. Owing to an HIV test kits supply shortage, this analysis excluded June and July 2010 data.
Kenyatta National Hospital/University of Nairobi Ethics and Research Committee gave ethical clearance (number KNH-ERC/A/474). Ministry of Public Health and Sanitation also gave written approval (Ref. No. MPHS/IB/1/14 Vol. III) to do the study and collaborate with Kenya National AIDS Control Program (NASCOP) and Department of Health Promotion. Nairobi Provincial Director of Public Health and Sanitation and the district head also gave written approval for the study to be carried out in Nairobi East district.
|
![[This page as RDF]](http://data.gesis.org/somesci/static/rdf-icon.gif){kind=icon}