Resources

Available information

Population-based estimates of the number of people using PrEP in a certain period (which can subsequently be used to create the numerator in estimates of 'PrEP coverage' and/or ‘PrEP-to-need ratio’). If unique identifier codes are available, longitudinal data can be used to estimate the number of PrEP initiations and/or develop indicators for PrEP continuation.

Benefits

1. Algorithms can be applied to specific databases to distinguish tenofovir disoproxil fumarate/emtricitabine (TDF/FTC) prescriptions for PrEP from other indications with high sensitivity and high specificity. For example, TDF/FTC for antiretroviral therapy (ART), hepatitis B virus (HBV) or post-exposure prophylaxis (PEP) infections.
2. Readily available data through routine monitoring (e.g. no additional data collection is needed).
3. Provides data at population-level.

Challenges

1. Databases often do not cover the entire population (e.g. no data on informal PrEP use or PrEP use registered in another administrative unit).
2. Data may not be representative of the entire PrEP-using population (e.g. in the context of missing differential data among non-insured individuals).
3. Databases frequently contain data on age, sex and postal code, but not on race/ethnicity or membership of key populations (e.g. MSM or sex workers).

Additional comments

Provides estimates on 'filled' prescriptions (proxy closer to actual use than written prescriptions).

Available information

Population-based estimates of PrEP use, HIV-testing adherence (as proxy for retention), and HIV seroconversions, through AIDSrelated virus (ARV) prescription or hospitalisation.

Benefits

1. Algorithms can be applied to specific databases to distinguish tenofovir disoproxil fumarate/emtricitabine (TDF/FTC) prescriptions for PrEP from other indications with high sensitivity and high specificity. For example, TDF/FTC for antiretroviral therapy (ART), hepatitis B virus (HBV) or post-exposure prophylaxis (PEP) infections.
2. Readily available data through routine monitoring (e.g. no additional data collection is needed).
3. Provides data at population-level.

Challenges

1. Databases often do not cover the entire population (e.g. no data on informal PrEP use or PrEP use registered in another administrative unit).
2. Data may not be representative of the entire PrEP-using population (e.g. in the context of missing differential data among non-insured individuals).
3. Databases frequently contain data on age, sex and postal code, but not on race/ethnicity or membership of key populations (e.g. MSM or sex workers).

Additional comments

Provides estimates on 'filled' prescriptions (proxy closer to actual use than written prescriptions).

Available information

1. Allows early investigation (preuptake) of cascade steps, such as 'awareness' or 'willingness-to-use'.
2. Allows investigation of relevant types of behaviour, such as – PrEP adherence, switches between dosing regimens, or HIV-risk behaviour (and hence PrEP eligibility).
3. Allows incorporation of relevant collections of sociodemographic data on PrEP users.

Benefits

1. Flexibility: questions can be adapted to fit local contexts.
2. Allows gathering data on individual knowledge, attitudes and types of behaviour that are often not addressed by other data sources.

Challenges

1. Large surveys are more likely to comprise convenience samples, and results may consequently not be generalisable to the whole population.
2. Self-reported outcomes are susceptible to information bias, including recall bias and social desirability bias.
3. Financial and human resources are required to develop, disseminate, administer and analyse surveys.
4. Possibility of low response rates.

Additional comments

1. Thus far, large-scale behavioural surveys to monitor PrEP have focused on MSM (e.g. EMIS-2017 survey).
2. Different sampling methods (e.g. venue, internet or telephone-based) may characteristically yield different population samples. Telephone surveys among the general population can be used to yield a representative study sample.
3. Generally, internet-based surveys are timely, have a lower cost than inperson surveys and have a broad geographical scope. Yet, attention should be paid to a possible digital divide.

Available information

Data collected at service-delivery sites for PrEP can be aggregated to provide national or sub-national estimates.

Benefits

1. The data is routinely collected as part of (clinical) records.
2. Possibility to collect client-level data on membership of key populations, PrEP regimen of choice, adherence and continuation.

Challenges

1. The burden of data collection is on the data providers.
2. There might be instances of missing data if the administrative load is high.
3. Requires streamlining of data collection across facilities to have meaningful data on a higher (e.g. national/regional) level. This needs digital reporting systems.

Additional comments

If digital information systems allow, clinically coded (client-level) data could be directly linked to a central database as part of routine surveillance.