Variables in the client table.

Regional district codes, division name, village name, client unique ID, date of birth, date first positive HIV test, date confirmed HIV-positive, referred from, marital status, drug allergies, community support group, prior exposure, transfer in ID, tuberculosis (TB) ID, home-based care, client code, height.

Variables in the visits table.

Client’s unique ID, visit date, weight, now pregnant, TB screening, antiretroviral (ARV) status code, ARV reason code, ARV code, CD4 count, CD4%, nutritional status, relevant co-medication, haemoglobin (Hb), height, WHO stage, functional status, ARV adherent, eligibility reason, visit type code.

Sample size estimation.

Sample size was calculated on a wealth index constructed from the economic data gathered, but the calculations could be applied to other factors in the questionnaire. We assumed that among the clients that come to a CTC within three months of receiving their first positive HIV test result, 40 percent are considered wealthy and 30 percent are not. To detect the difference of 10 percent with 80 percent power and 5 percent significance, a simple random sample of 750 clients was needed. With the expected clustering in clinics, we assumed a design effect of two. Therefore, a sample size of 1500 clients was calculated to achieve 80 percent power at 5 percent significance to detect a 10 percent difference in the proportion of those who attend CTC within three months of their first HIV diagnosis.

We had no predetermined sample size for a qualitative part of this study, participants were recruited until when the saturation point was reached.

Sampling method.

A list of health facilities considered to have a high volume of clients was first identified in each district. Then a simple random sampling selected five CTCs per district. Priority of selection was given to the CTCs considered to have a high volume of clients and then to the health facilities close to having a high volume of clients. Following that, a total of 15 CTCs from three districts (five per district) were identified. From each selected CTC, at least 50 attendees who met the study inclusion criteria during the survey period (two months) were invited to participate. Then everyday five clients were selected randomly from the daily client’s appointment logbook. The same process was repeated every day until the planned 50 clients per clinic were enrolled. While for qualitative part, a convenience method was used for recruiting study participants for the FGDs and IDIs.

Data management and analysis.

Qualitative data were analysed using a thematic analysis framework with six main steps: familiarizing oneself with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report (Braun & Clarke, 2006). To capitalize on validity and reliability, the collected information was given a code for confidentiality. Field notes were expanded into full sentences and paragraphs by adding details of what was observed using the notetaker’s notes and audio recordings. The expanded notes were translated into English, typed in Microsoft Word 2007, and stored on a computer. Narratives were read, coded, and cross-checked. Six main themes emerged from the data: referral letter/forms, escorting clients, stigma, falling sick, denial of results, and counselling procedures.

Data collected through quantitative methods were entered in an Access database (interfaced by Epi Info3.5.4). The database was developed to correspond to the study questionnaire. The data were cleaned for quality control purposes and transferred into Stata format to form a primary data set. The data from the CTC2 databases were extracted based on the study variables and transferred into Stata format to form a secondary data set. The two data sets (the one obtained from primary data collection and the other from the CTC2 database) were then merged using a unique CTC ID number to form a single data set.

The outcome of interest in this study was whether a client enrolled at a CTC within three months of the date of his or her first HIV-positive test. The period was calculated by subtracting the date of registration at a CTC from the date of the client’s first positive HIV test. Descriptive analyses were used to capture different characteristics of the study participants. Continuous variables were described in terms of median and interquartile range, and categorical variables were described in terms of frequencies and percentages.

World Health Organization (WHO) clinical stages and CD4 counts were used to document the disease progression between HIV clients presenting late at CTCs and those presenting early. The changes in baseline and follow-up were determined for both the WHO clinical stages and CD4 counts, and these were compared between the groups. In documenting the disease progression, the proportion of HIV clients in advanced clinical stages (i.e., WHO clinical stages III and IV) was determined in both groups (i.e., for those presenting late at CTCs and those presenting early) at baseline, six months, and 12 months. Two kinds of analyses were carried out to describe the disease progression, in terms of CD4 counts, for the two groups. The first analysis examined the disease progression based on the proportion of HIV clients who have CD4 counts of less than 350 at baseline, six months, and 12 months and compared the two groups, and the second analysis compared the disease progression in the two groups in terms of median CD4 counts at baseline, six months, and 12 months. The analysis was carried out to HIV clients in the study districts who had both WHO clinical stages and CD4 counts at baseline, 6month and 12months.

Cross tabulations and crude odds ratios were conducted to assess preliminary associations between independent variables and the outcome of interest. Bivariate analysis assessed potential confounders between variables that were statistically significant in the univariate analyses. The relevant confounders were ascertained when the estimate/odds ratio changed by 20 percent or more. Following the assessment, no relevant confounders were identified. Multivariable logistic regressions were conducted for the factors that were statistically significant from the univariate analyses. The final model was chosen by performing the stepwise model selection by Akaike Information Criterion (AIC). And for time changing variables including age, marital status, and level of education, the baseline data was used. The clustering effect was taken into consideration to adjust for variations within and between clinics. All the analyses were carried out using STATA Version 12 with the results summarized and presented in tables and figures.

Factors associated with effective linkages of newly diagnosed HIV clients to CTCs.

All participants in the study were aware of HIV-related information on linkage strategies and referral systems. Participants’ responses were consistent with many of the factors explained by participants to be barriers to effective linkages and referrals repeated in the FGDs and IDIs across the study sites.

Referrals.

Referral forms, or referral letters as some called them, were noted by all FGD participants as being the main method used to link new clients to CTCs. According to their responses, this form/letter was designed by the Tanzanian Ministry of Health Community Gender Elder and Children (MOHCGEC) and sent to all health facilities that have a CTC. After a client is found to be HIV-positive, the form is filled in and addressed to the referral health facility chosen by the client. A respondent from Bombo.

Regional Hospital said:

The method used is still the same. After a patient tests positive and he or she decides which VCT centre to get treatment, let’s say he or she says, ‘I want to be treated at Bombo CTC,’ so you must give him the letter.

Another respondent said, before this form is filled out, a client must be asked where he or she comes from so that the client can be easily linked to the facility nearest his or her area of residence.

First of all, you ask the client that has come to you; you ask him or her where he or she lives, because nowadays HIV medications are available in many centres; you need to ask where that patient lives. He or she tells you, I live in a certain area, and then you mention the CTC which are within his locality.

There was an insistence on knowing where a client comes from, to facilitate the movement of the client, while accessing care. One respondent from Ngamiani Health Centre said,

A client goes to VCT; he or she gets tested there. That client came from home and went to get tested; he or she is given a form and explained where he or she is supposed to go, okay? Because at VCT, a client may come to Ngamiani but he or she is from a different district, so he or she will be given a form and asked to go back to where he or she came from.

It was also reported that at times there are linkages based on departments within a health facility that link a client from one point to another, based on the nature of the disease found.

Now for example at the RCH, they have a special form which they use; when a patient is diagnosed with HIV, there is a form which is filled and given to a patient to take to CTC.

However, some linkages for newly identified clients are from people other than healthcare workers and do not involve referral forms. One such client reported being linked by a friend. Another participant, from Bombo Regional Hospital, reported,

It is my mother who linked me with a CTC because she is also HIV-positive, and I was born with HIV.

She continued,

To be honest, at first, because I work at a hospital, first I tested myself and saw the results. Immediately you know that situation, I mean I got worried. After getting worried, my colleagues counselled me and I came here at CTC to start medication.

Many said a problem with the referral forms is that they lack important information that would allow healthcare workers to follow up easily with clients who have tested HIV-positive. They suggested that the MOHCGEC modify the form to include other important information that can be helpful when looking for an HIV-positive client when he or she does not comply with attending CTC.

Escorting clients.

Escorting clients was one of the ways proposed in the FGDs to link new clients to care and treatment. According to the respondents, this is done based on the client’s level of understanding and is mostly used when a client has chosen to get care within the same healthy facility:

Every time you explain to him or her to come to CTC and he or she does not understand, maybe he or she asks to be escorted, then you can decide to take him or her; but in most cases, we just give them the referral letters, and they go to the CTC of their choice.

Escorting new clients came about because clients were not attending the CTC to which they had been referred. One respondent said,

At first, we saw that there was a problem: that clients did not go to the CTC which they were referred to. So, we used that method (escorting), especially those who were from the reproductive and child health clinic.

During FGDs respondents expressed that escorting a client is fine if the client requests it, but an escort is required if the client is unconscious. For IDIs, most of the respondents thought an escort was the best referral mechanism, because most of them did not want their status to be easily disclosed to other people—even family members who may not know they are HIV-positive or even that they are on ART. The common suggestion was for health service staff to be available always so clients can choose where to get services.

Stigma.

Procedurally, whenever a client tests HIV-positive, there must be a referral to the CTC nearest his or her place of residence, to make it easy for him or her to get treatment. However, even with this protocol in place, some clients have not been attending, or have delayed attendance at, a CTC due to various factors such as stigma. Stigma was evident both internally (with self-stigma, by the client) and externally (with stigma from relatives, healthcare workers, and the community). An FGD respondent from Handeni Hospital said:

Most who come to get tested are convinced by their friends or relatives. They say, ‘let’s go get tested; let’s go get tested.’ When he or she is tested, and found to be HIV-positive, he or she is with them, this person is not ready to let the colleagues that came with him or her to know his or her status at that time, so even if you give him or her the letter, he or she will not come.

Another FGD respondent expressed that clients may be worried about stigma from their relatives, which creates a delay in seeking treatment:

One may think ‘if I disclose my status, how will my relatives think of me? How will I start telling them that I did go the hospital?

According to a respondent, stigma from the community also presents a problem:

Another factor could be fear of how the society will treat it based on the reality that he or she is infected with HIV, when one thinks about the stigma that may come from the relatives or family, he or she may decide not to disclose his or her status.

Because of stigma, most clients prefer to travel a longer distance to obtain their care and treatment than visit their nearest facility to avoid the risk of encountering people who might recognize them and find out their HIV status.

Falling sick.

Falling sick was brought up by a respondent as a factor that delays new clients from attending a CTC after they are referred:

It is the short-sighted mind that I mentioned. A person may decide, to wait and see what will happen. Will I get sick or not? If they don’t get sick, they believe they are not infected but if they get sick, then they believe that they are really infected with HIV.

The habit of waiting to fall sick before seeking treatment is often a fatal decision, because the person’s health has already deteriorated by the time the individual reaches the hospital. One respondent explained,

[They give themselves] hope that, maybe it is not true; let me wait and see how it goes. Because they delay too much in seeking care, that is why by the time they come, they are really worn out (really sick). They say, let me wait and see the end, what is the end? They come to the hospital too sick that some cannot even walk. They have to be assisted by their relatives.

The decision to wait until the person falls sick is related to other reasons that respondents gave for putting off seeking treatment, such as denial of results and feeling well even after testing HIV-positive. By the time these clients accept their health condition, it is often too late for treatment. The discussion also revealed that most of the clients do not understand the consequences of starting treatment late. If they knew the severity of that decision, presumably, fewer would wait to get treatment.

Denial of results.

According to respondents, most clients delay following through on their referral to a CTC, because they are in denial of their HIV-positive test results. Emotionally, they may not be able to face an HIV diagnosis, or they may believe that the results given to them were false. One informant said,

Some patients come to test for HIV out of planning, someone sits down and decides to go and test and if they find they are positive they run away and reject the results given to them. At the end of the day they delay attending to CTC for more treatment.

Likewise, an IDI respondent explained,

Some maybe don’t believe they are infected. They do not believe that they actually have HIV, and some may think, if I have HIV, why should I go to the hospital before infecting other people.

Another said,

The client may not have expected to be diagnosed with HIV, and most who come for tests, do not expect to be diagnosed HIV-positive. So, when some test and are diagnosed HIV-positive, some understand and accept it, but for some, it is not that easy. That is why some would need to go back home and think about it (their newly diagnosed HIV-positive status), and they tell you for now, even if you have given me a referral letter, let me go home first and think about it, prepare myself, and see if I am ready to start CTC or not. So, for many it is how they receive the results which makes them not come directly to CTC but instead go home first.

Counselling Techniques.

Knowledge and competency of the counselling officer could also contribute to delays, as some may not clearly convey the consequences of failing to attend a CTC after the client has tested positive for HIV.

It depends on the counsellor that he or she will meet with. If he or she meets a counsellor who is not competent, he or she may not focus a lot on telling the patient to go to CTC, because he or she looks healthy; he or she does not have any symptoms of HIV infection. He or she may thus wait until he or she gets symptoms before going to CTC.

Most of the PLHIV said that some healthcare providers do not provide detailed counselling. Instead, they give the results and ask for the nearest health facility, to address the referral form. This is done as a formality and is sometimes a result of having too many clients to attend to. Another respondent said that they are given group counselling, without consideration for individual needs and preferences.