Title:Contraceptive use and Unmet need for Family Planning among HIV Positive Women on Antiretroviral Therapy in Kumasi, Ghana

Transcription

1 Author's response to reviews Title:Contraceptive use and Unmet need for Family Planning among HIV Positive Women on Antiretroviral Therapy in Kumasi, Ghana Authors: DENNIS O LARYEA (dlaryea@kathhsp.org) KATHRYN SPANGENBERG (k_spangenberg@yahoo.com) YAW A AMOAKO (yamoako2002@yahoo.co.uk) EBENEZER FRIMPONG (kolobo24@yahoo.com) Version:2Date:16 June 2014 Author's response to reviews: see over

2 16 th June, 2014 Dear Sir/Madam MS: Contraceptive use and Unmet need for Family Planning among HIV Positive Women on Antiretroviral Therapy in Kumasi, Ghana We have revised our manuscript as recommended by the reviewers and wish to submit it for further review. Reviewer 1- Oskari Heikinheimo The authors have examined the patterns of contraceptive use among female attendees of an antiretroviral therapy clinic in Kumasi Ghana. The authors find that less than half of the couples were using contraception, the use of modern contraceptive methods is rare and that there is a high unmet need of contraceptive services. Even though the results come as no surprise, I applaud the authors for their efforts in studying a very important local & national issue. Although the data is of interest, the study is a descriptive one, the population studied (n=230) comes from a single clinic and represents only a small fraction of all HIV-infected fertile aged women in Ghana (total nro of HIV-infected individuals approx ). I therefore think that the paper would be better suited for a national / local journal. The reviewer makes no specific requests for amendments to the article. On the issue of our study population representing a small fraction of HIV positive persons in Ghana, while we agree that the sample in this study is small, we know that only about 59,000 persons are on ART in Ghana including children. Our study was clear on the target population for the study which is HIV positive women on ART hence it is not as negligible a proportion. We believe also that this will stimulate further research in the general population to include persons living with HIV but not on HAART Reviewer 2: Frank Kaharuza Major Compulsory Revisions 1. Background Dual protection and dual methods for HIV infected women is a recommended practice that should be described for this paper. This will help the readers understand the critical role of condom use among the HIV infected individuals and perhaps explain the high condom use We have included in our literature review added some information on dual protection by inserting the following: Modern methods of contraception such as sterilisation, intrauterine devices, male and female condoms among others are effective means of preventing unwanted

3 pregnancies. Correct and consistent condom use confers dual protection (preventing unwanted pregnancy and sexually transmitted infection) when used alone or with another non-barrier method of contraception although some evidence suggest this practice may not be common (12 14) Methods The methods section lack information on the data collected- and how it was managed- the authors should tell us what data was collected- and how it was manipulated Results We have added the data collection tool and the method of manipulation with the insertion of the following sentences in the section on methods: The data collection instrument was a structured questionnaire administered by two trained research assistants. The broad areas covered by the questionnaire were demographic information, history of sexually transmitted infections, fertility desire and contraception usage. Following written informed consent, all eligible consenting respondents were interviewed between December 2012 and March 2013 Data collected data was entered into a Microsoft Access Database and. Main outcome: contraceptive use: It is important to show what the contraceptive method mix is for all the participants. Table 8 only shows the method mix for those that no fertility desire. Unless the readers understand the method mix in this population it will be assess contraceptive prevalence We indicated the various methods used by all respondents in the Results section with the sub-heading Contraceptive use. We also presented this in figure 1. Are the contraceptive methods used mutually exclusive. In the HIV population we need to understand if dual methods are being used- and are they also being promoted. The prevalence of Dual method use should also be presented if it was collected. Figure 1 presents the distribution of methods used but it will be important to include dual methods We highlight the reason we did not include dual methods as both contraceptive prevalence and unmet needs for family planning consider only one partner using a method irrespective of dual use. We however recognise that dual protection is a major issue in HIV contraceptive use. Reported Partner desire for children.: The author should clarify respondent and partner desire. There are four potential outcome: Both Partner and respondent desire children, Partner disires respondent doesn t, respondent desires partner doesn t, and both don t desire- currently you

4 report 53.5 respondent desire children 54.6% partners desire children- but no information on discordant disires We have added the proportions as advised by inserting a table detailing the various combinations of fertility desires among women with partners in this study. We also inserted the following: Among respondents with a partner (n=142), the desire of both partners to have a child was found among 48.6% of respondents. The details of the various combinations of fertility desires among respondents with a partner are as shown in table 4. Desire for a child Frequency Percentage Both desire Woman alone Partner alone None desires Total Limitations Please clarify that partner use is reported by the respondent. It is therefore important to consider this as a limitation of the study- since condom use is recommended in HIV positive individual and the response on condom use could be high because of social desirability We have inserted the partner use of a condom as reported by respondents as a limitation by inserting: Partner use of condoms was as reported by respondents and may have been influenced by the social desirability for condom use by HIV positive persons. Did the authors collect information on education? This would have been desirable information to have since education has a role to play in uptake of contraceptive services. If it wasn t collected then this should be mentioned in the limitations We have added information on the educational background of all respondents in the study by appending the table below to table 1:

5 Educational Status Nil Primary JHS 1 /Middle School Secondary Tertiary We also analysed for trends in uptake relative to educational status using chi-square test for trend and included the result of this analysis in the results section by inserting: The educational status was associated with an increased likelihood of contraceptive use on chi square test for trend from respondents with no education to those with tertiary education (OR= 1.43). This association was however found not to be statistically significant (p= 0.23). Any comment on the wide confidence intervals seen in the results could be included given the small sample size We have inserted the observation of wide confidence intervals as being possibly due to the small sample size. We inserted the sentence below in the section on limitations: The wide confidence interval observed for the association between previous contraceptive use and current usage may be due to the small sample size in this study. Minor Essential Revisions The author can be trusted to make these. For example, missing labels on figures, the wrong use of a term, spelling mistakes. Methods We have checked to ensure that all tables and figures are correctly labelled. We have also checked the usage of terms in the manuscript. 1 Junior High School

6 Using a design effect of 1 is similar to no design effect? Since you used SRS, you don t need to include the design effect in the sentence We have deleted the phrase and a design effect of 1.0 The last sentence of Paragraph 2 in the Methods Section now reads: Based on Ghana s current contraceptive prevalence rate of 25.2%, a sample size of 229 was estimated using Epi Info version at a 95% confidence interval. How did you define previous STI that would give you a high STI prevalence among respondents? Please include in the methods section We made reference in the results section, paragraph 3, on the parameters looked out for with reference to STIs i.e....sexually transmitted infection (STI) in the form of a genital ulcer or discharge in the one year preceding the study Tables: Table structure: there are many tables presented and some of the tables can be merged for example tables 5,6,7 present data on association of marital status and contraceptive use We have merged tables 6 and 7. We maintained Table 5 because there were fewer columns compared with the other two although all three tables relate to marital status as noted by the reviewer. We now have Table 7 as a merger of tables 6 and 7. Results In text results for clarity it may be necessary to state the denominator of interest in presenting some of these results- for example. Among those with current sexual partners (PROVIDE THE N), only 3.2% of the respondents indicated a partner STI during the one year We have indicated the total number in bracket and also indicated the number of respondents who could not provide an answer for this question by inserting: Among those with current sexual partners (n=125), only 3.2% of respondents indicated a partner STI during the one year preceding the study. A total of 17 respondents with partners could not provide information on partner STI in the preceding year. Discretionary Revisions Methods The sentence on ethics can be revised to improve readability data collection was between December 2012 and March 2013 and Participation was voluntary and followed the receipt of informed consent

7 We have amended the sentences and it now reads: Following written informed consent, all eligible consenting respondents were interviewed between December 2012 and March Discussion you opinion on the mean age and We have added the following to the first paragraph of the Discussion section: The recorded mean age of 36.3 years and the modal age group of years. This has implications for attempts at reducing MTCT as these women are in their reproductive ages(25,26) Reviewer 3: Fredrick Makumbi Major Compulsory Revisions Abstract should have a brief but clear method section We have amended the section on the methods and now reads: This was a descriptive cross sectional study of HIV positive women in their reproductive ages accessing care at an adult Antiretroviral Therapy Clinic in Kumasi, Ghana. Data was collected using a structured questionnaire. Data analysis was conducted using Epi Info version It is not very clear, if this was any contraceptive use or only modern methods? It is important that this comes out clearly We have clarified this in the abstract by inserting the following to the introduction: We set out to measure the prevalence of modern methods of contraception, the unmet need for family planning and to identify factors associated with the use of modern methods of contraception among HIV positive women on ART. I also strongly suggest that your analysis has use of methods categorized as 1) None 2) Condoms only 3) at least a modern method. This is very critical because your study population is HIV+ and may be using condoms (whose frequency and consistence you do not know) for HIV co-infection/infection of their partners. Thus knowing use of more effective methods is important. We do acknowledge the reviewers suggestion that the analysis focus on the three parameters he gave. However, we state in our objective that we intend to assess the unmet need for contraception and contraceptive prevalence and these do not require the categorisation of the methods but rather the use or otherwise. In addition, the effectiveness of modern methods of contraception is well known (both effective and

8 typical use). We have indicated in our paper, the various methods being used by respondents in order to be able to have a good idea of the contraceptive use and the expected effectiveness. The conclusion should reflect the key reasons you set our to undertake the study. Your abstract only makes conclusion on the unmet need, yet it was not the only key objective you set out to conduct the study We have amended the conclusion to the abstract and added: There is a high prevalence of contraceptive use among HIV positive women in this study. However, the... Background: Please ensure that facts from your literature review are well summarized and stated. Sentence one has a lot of miscommunication of facts! We know that HAART has not reduced the prevalence of HIV! Rather HAART has improved survival of the persons, and reduced frequency of opportunistic infections. Please restate this first sentence We have amended the first sentence and it now reads: The advent of Highly Active Antiretroviral Therapy (HAART) has resulted in significant improvements in the health of persons infected with the Human Immunodeficiency Virus (HIV) [1]. You need to state that women on HAART are at increased of conception because of improved physical health that may lead to more frequent sexual intercourse, and improved immunity. These facts are missing in your background yet they form an important background to why you need to study use of contraceptive among HIV+ women We did not find evidence in our literature review that women on HAART are more likely to engage in sex although it makes sense when comparing persons with advanced HIV infection with the general population. We relate the relevance of this study on the need to institute measures to increase the use of contraceptives among HIV positive women to reduce the unmet need for contraception and contribute to achieving the target for eliminating MTCT of HIV The whole background should be revised so as to streamline it, be coherent and provide only relevant information that informs your intended study. So please in a coherent manner clearly state what the problem is, its magnitude, its consequences, what is already known from previous studies, what are the gaps in knowledge and so what you will address(objectives). These questions may help guide you in what key information to include in your background, but in a coherent manner We have revised the background section. The following have been deleted:

9 As a result, persons infected with HIV are living longer and are thus likely to be exposed to challenges with fertility and contraception as occurs in the general population. Despite some challenges particularly in the area of funding, the programme has made some gains in the management of HIV infections in Ghana. These include increased access to ART from about a few hundreds to 59,000 at the end of 2011[5] High fertility rates have implications for both maternal and child health. Factors such as age, marital status and the death of a child affect fertility rates [8]. Uncontrolled population growth has both social and economic implications. The global population has doubled in the last 61 years [9]. By comparison, Ghana s population has almost quadrupled between 1961 and 2010 with an estimated 6.7 million population in 1961 increasing to 24.7 million in 2010 [10 12]. High variations in the proportions of unintended pregnancies have been reported. There has been a general decline in fertility in sub-saharan Africa over the years albeit with some variation in rates [24,25]. Female education, decline in child mortality, urbanisation and modern culture have been identified as some of the factors accounting for a decline in fertility rates in some parts of Africa [25]. This notwithstanding, fertility rates are still high. The World Health Organisation (WHO) and the United Nations Children s Fund (UNICEF) seek to achieve this through the prevention of unwanted pregnancies among women with HIV and the provision of ART to all pregnant HIV positive mothers and safer options for delivery and breastfeeding. We added the following to the introduction: Modern methods of contraception such as sterilisation, intrauterine devices, male and female condoms among others are effective means of preventing unwanted pregnancies. Correct and consistent condom use confers dual protection (preventing unwanted pregnancy and sexually transmitted infection) when used alone or with another non-barrier method of contraception although some evidence suggests that practice of dual contraceptive use may not be common (13 15). We set out to estimate the contraceptive prevalence rate and the unmet need for family planning among HIV positive women in their reproductive ages on HAART in Kumasi and to examine factors associated with the use of modern methods of contraception in this population.

10 Methods; Please provide more clarity in write up on inclusion criteria, how eligible women were identified (see suggested write up in the submitted manuscript). We indicated the basis for selecting respondents in the second paragraph of the section on methods. You need to justify why you use a design effect, and why it should be 1 We included the design effect by error as it makes no difference mentioning a design effect of 1 so that has been deleted. There is almost no description of the data analysis, which is a very important section of any scientific writing! We have provided more information on the data analysis by inserting the following: Data collected was entered into a Microsoft Access Database. Epi Info version was used to calculate frequencies and proportions. Chi-squared tests were conducted for variables with categorical or binary outcomes. It is not clear how the results were obtained without analysis plan We have indicated in the preceding section how data was managed. It is not stated if the OR were crude or adjusted. You need to state this. However, the final analysis from where you draw conclusions should have adjusted measures of associations, and presented as adj.or We have indicated the OR used as an appendage to tables 5 and 7 by adding: *Unadjusted OR Results of unmet need are so scanty and the definition of unmet need was not provided. These must come out very clearly. We have included the definition of unmet need for contraception in our background information by inserting The WHO defines unmet for family planning as Women with unmet need are those who are fecund and sexually active but are not using any method of contraception, and report not wanting any more children or wanting to delay the next child (12). Figure 2 presents information on the estimation of the unmet need for family planning. Other information provided in the results such as Table 3 and 4 which

11 includes information on desire for children and current use of a modern method of contraception. A key variable in the analysis is duration on HAART in relation to these outcomes, i.e. unmet need and contraceptive use. Can you please have this done? We have added a table on the duration of HAART which we categorised into 5 and conducted a logistic regression. Table 1: Duration of ART use among respondents Duration of ART use Frequency Percent 12 months months months months months >60 months The duration of ART use was also not associated with contraceptive use (table Table 2: Duration of ART and contraceptive use Parameter Odds Ratio 95% C.I. P-Value 12-23mths/<12mths mths/<12mths mths/<12mths mths/<12mths >59mths/<12mths The discussion ; It seems to only repeat results, again without much explaining the why and comparing with previous studies. You should be able to compare, contrast and explain your findings in the context of Ghana or specifically Kumansi. For example you need to explain why condoms are the most commonly used methods, and what that means for prevention of pregnancies. What about other methods? What is the level of use of those other methods? We have amended the section by adding the following to paragraph 2 of the discussion section.

12 Condoms may be the commonly used method among the respondents in this study because of the conferment of dual protection, its ready availability, relative ease of use, promotion by the health authorities in Ghana and the avoidance or medication and perceived side effects of other contraceptives. Consistent condom use has been associated with an 80% reduction in HIV incidence(32) The use of other modern methods of contraception among the respondents in this study was low (20.4%- dermal implants, oral pills and Injectables) compared with condoms (79.6%). While a high contraceptive prevalence rate is generally desirable, the use of condoms in particular is important among HIV positive women particularly in cases of discordance. The low use of other modern methods compared with condoms among respondents in this study is therefore an important finding. What does it mean if they are low, what would you recommend? Is the low use of other methods a consequence of availability of FP in this setting? We add the need for integrating FP services in our discussion. The following has been inserted in paragraph 3: Family planning services are not available at the Adult HIV clinic and this may have been a factor in our findings in this study. There may be the need to integrate family planning services in the Adult HIV clinic in Kumasi A clear example is a desire for children? if you have 50% in Ghana desiring for children,yet Uganda and SA is about 31% what does this mean? Explain these findings and what implications it may have in Ghana We add the possible implications for the high fertility desire among women in this study by inserting into paragraph 4 of the section on discussion The high fertility desires recorded in this study may also have implications for the elimination of MTCT of HIV programmes in Ghana. Women on HAART may be at a lower risk of transmitting infection to their children but a high fertility rate means the number of HIV positive infants may also increase as a result. Limitation; I think the limitation are not well explained and do not fit to be limitation. Inclusion of infecund women in the analysis for contraceptive use and unmet need can be handled. The re-analysis should be able to exclude these women. Please think of other limitation This study could not determine the fecundity or otherwise of women in this study hence our decision to highlight this as a limitation. We will therefore be unable to determine these women and exclude them in our analysis. While contraceptive prevalence focuses purely on the use of a modern method of contraception with no emphasis on ability to reproduce, unmet need for contraception does hence we cannot also conduct such analysis in relation to contraceptive prevalence excluding infecund women

13 The first limitation indicates a 34% CPR relative to 25% in general population. However, the results section indicated 42% CPR. We have deleted the following from the section on limitations as the issue of pregnancy applies more to unmet need for family planning than contraceptive prevalence. We however believe that the difference between our recorded contraceptive prevalence of 34% compared with a national prevalence of 25% may be attributable to the exclusion of this group of women. The conclusion represents findings, however, the findings of CPR are driven by use of condoms in this HIV+ population. This can be very miss leading. I strong suggest that you conduct analysis of other modern methods alone Condoms are considered modern methods of contraception and are included in the estimation of the contraceptive prevalence rate. Our exclusion of condoms in the estimation of the contraceptive prevalence rate will overly underestimate the contraceptive prevalence rate. The usefulness of condoms as a method of contraception has also been highlighted throughout our text. Minor Essential Revisions - Minor Essential Revisions There is a general need to write in a more clear and coherent scientific writing style Amendments have been made in the preceding sections to improve on the clarity and coherence of the text. Statistical presentation may only require the measures of association, 95% CI, and may be p-values. Note that P-values that are too small, e.g. p= can be presented as p<0.001 We have amended the results in the tables and in the body text of the p-values as shown below Table 5: Predictors of contraceptive use among HIV positive women in Kumasi Parameter Odds Ratio 95% CI p Desire to have children Previous STI Partner with previous STI Partner desire for a child Previous contraceptive use before <0.001 diagnosis Partner knowledge of HIV status <0.001

14 No need to present chi-square, especially if there are no degrees of freedom because thy can not be interpreted. Chi square values have been deleted from all results We look forward to a favourable response. Thank you Yours Sincerely DOL, YAA, KS, EF and JKA