**Clinical Utility of Delta Lactate for Predicting Early In-Hospital Mortality in Adult Patients: A Prospective, Multicentric, Cohort Study**

**Pablo del Brio-Ibañez 1,**†**, Raúl López-Izquierdo 2,**†**, Francisco Martín-Rodríguez 3,\*, Alicia Mohedano-Moriano 4, Begoña Polonio-López 4, Clara Maestre-Miquel 4, Antonio Viñuela 4, Carlos Durantez-Fernández 4, Miguel Á. Castro Villamor <sup>5</sup> and José L. Martín-Conty <sup>4</sup>**


Received: 19 October 2020; Accepted: 16 November 2020; Published: 17 November 2020

**Abstract:** One of the challenges in the emergency department (ED) is the early identification of patients with a higher risk of clinical deterioration. The objective is to evaluate the prognostic capacity of ΔLA (correlation between prehospital lactate (pLA) and hospital lactate (hLA)) with respect to in-hospital two day mortality. We conducted a pragmatic, multicentric, prospective and blinded-endpoint study in adults who consecutively attended and were transported in advanced life support with high priority from the scene to the ED. The corresponding area under the receiver operating characteristics curve (AUROC) was obtained for each of the outcomes. In total, 1341 cases met the inclusion criteria. The median age was 71 years (interquartile range: 54–83 years), with 38.9% (521 cases) females. The total 2 day mortality included 106 patients (7.9%). The prognostic precision for the 2 day mortality of pLA and hLA was good, with an AUROC of 0.800 (95% CI: 0.74–0.85; *p* < 0.001) and 0.819 (95% CI: 0.76–0.86; *p* < 0.001), respectively. Of all patients, 31.5% (422 cases) had an ΔLA with a decrease of <10%, of which a total of 66 patients (15.6%) died. A lactate clearance ≥ 10% is associated with a lower risk of death in the ED, and this value could potentially be used as a guide to determine if a severely injured patient is improving in response to the established treatment.

**Keywords:** prognosis; lactate clearance; biomarker; emergency medical services; emergency department; critical care

## **1. Introduction**

One of the challenges in the emergency department (ED) is the rapid identification of those patients who, upon arrival, may have a greater risk of clinical deterioration, which may lead to serious adverse events (SAE), such as unplanned admission to the intensive care unit (ICU), major adverse cardiovascular events, or early mortality [1].

Although there are a series of early warning scores based on different physiological parameters, which are capable of predicting the risk of deterioration in EDs [2,3], there are still situations in which SAEs could be detected earlier if there was an effective early warning [4].

Therefore, different biomarkers with prognostic value are being evaluated, such as lactate [5]. Under normal physiological conditions, lactate production remains constant with lactate consumption; prolonged hyperlactacidemia (serum concentrations > 4 mmol/L) is the result of an increase in production or a reduction in consumption [6]. Hyperlactacidemia is often caused by an imbalance between oxygen supply and demand, and therefore elevated lactate can be seen as a non-specific marker of tissue hypoxemia, with this being a documented risk factor for mortality in patients with a serious and, more specifically, an infectious pathology [7,8].

The predictive value of a single lactate measurement as an indicator of hypoxic cellular distress is being investigated [9], and even more so, to detect mortality beyond the first 24 h [10]. A second lactate measurement can help to quantify the change from the initial measurement, which is called delta lactate (ΔLA), with a direct relationship with mortality [11,12].

The measurement of lactate levels in the ED is a routine analytical procedure [13], and point-of-care testing is beginning to be implemented in emergency medical services (EMS) [14]. Therefore, at this time we have a high level of evidence of the prognostic value of lactate, both in the ED and in the prehospital setting [5,15,16].

The primary objective of this study was to evaluate the prognostic capacity of ΔLA (correlation between prehospital lactate (pLA) and hospital lactate (hLA)) with respect to early in-hospital mortality (up to two days from the index event). The secondary objective was to analyse the predictive capacity of ΔLA for 7 and 30 day in-hospital mortality.

## **2. Experimental Section**

## *2.1. Study Design and Setting*

We conducted a pragmatic, multicentric, prospective and blinded-endpoint study in adults who consecutively attended and were transported in advanced life support (ALS) with high priority from the scene to the ED between the 1 October 2018, and 30 November 2019.

The study was carried out by six ALSs who transferred patients to five hospitals of the public health system (Burgos University Hospital, Segovia Hospital Complex, Salamanca University Assistance Complex, Rio Hortega University Hospital and Valladolid University Clinic), with a reference population of 1,351,962 inhabitants.

EMS operates non-stop 24/7 every day. Requests for assistance are evaluated by a physician at the emergency coordination centre who determines the most appropriate resource based on care needs. The ALS is made up of a physician, an emergency registered nurse (ERN) and two emergency technicians. On the scene or en route, they perform standard advanced life support actions according to the protocols for each pathology. Patients are transferred by the ambulance team to the ED. In the triage area, an ERN determines the level of priority and then hospital care begins.

This study was approved by the Research Ethics Committee of all participating centres (reference REC: #PI 18-010, #PI 18-895, #PI 2018-10/119 and #CEIC 2049) dated 9 March 2018. The study protocol is available online (doi.org/10.1186/ISRCTN17676798); we follow the STROBE guidelines for reporting. All patients (or guardians) signed the informed consent, including consent to data sharing. The ERN of the ALS attempted to obtain informed consent. If the patient's clinical situation or level of consciousness did not allow this, an ED physician tried again to obtain consent. In situations, such as death, or patients referred to the ICU in which it was not possible to obtain the document, a relative or legal guardian was contacted to ensure that informed consent was obtained.

## *2.2. Selection of Participants*

A patient was considered to meet the criteria to be included in the study if they had been evaluated and transferred by an ALS to the ED of the referral hospital and did not meet any exclusion criteria, among which are: under 18 years of age, presence of cardiorespiratory arrest, death prior to or during transport, pregnant women, patients with an acute psychiatric pathology or those with a documented terminal illness. Those which were also excluded from the initial cohort were those who, even meeting the inclusion criteria, had not undergone a hospital lactate analysis or those who had not been able to complete follow-up, due to lack of data or duplication. If a patient was admitted more than once during the study period, only the first admission was counted. In cases in which informed consent was not obtained despite multiple attempts, the case was excluded.

## *2.3. Outcome Measures and Study Protocol*

The main outcome variable was in-hospital mortality within 48 h from any cause, and secondary in-hospital mortality at 7 and 30 days was also analysed.

## *2.4. Study Protocol and Collection of the Parameters*

A procedure was developed for the determination of pLA, the operation of the equipment, cleaning, maintenance and calibration and specific training was carried out for all members of the EMS. The traceability of all the test strips used in the study has been monitored, by checking the expiration, serial number and lot number.

For the data collection, a standardised form was designed (medical history routinely used by EMS), where the ALS physician recorded demographic variables (age and gender), standard vital signs and prospectively the pLA value. All the prehospital clinical data analysed refer to the team's first contact with each of the patients. In the ambulance or on the scene, a venous blood sample was obtained with which pLA was determined. The analysis was performed using the Accutrend® Plus meter (Roche Diagnostics, Mannheim, Germany). All the measuring devices were calibrated every 100 determinations, always by the same researcher from each ALS, using Accutrend® BM-Control-Lactate control solution (Roche Diagnostics, Mannheim, Germany).

During the first hour of ED care, a new blood test was performed on those patients who required it, and hLA was determined together with the rest of the standard analytical parameters. Thirty days after the index event, an associate researcher from each hospital, by reviewing the electronic medical record (JIMENA-SACYL), the hospital outcomes were obtained: hLA value, need for admission and/or ICU, data from 2, 7 and 30 day in-hospital mortality, days of admission and diagnosis.

With the two lactate measurements, clearance was calculated according to the usual formula for the established time [17–19],

$$\text{Lactate clearance} \left( \% \right) = \frac{\text{initial latitude} - \text{Follow} - \text{up factor}}{\text{Initial factor}} \times 100 \tag{1}$$

## *2.5. Statistical Analysis*

The database was designed and organised after the collection of double-entry data in order to reduce transcription errors. To guarantee the correct traceability of patients between the prehospital setting and hospital care, the link criteria between the EMS history and the hospital electronic history were the date, ALS code, time of arrival at the ED, patient affiliation, gender and age. Prior to statistical analysis, the database was cleaned using logical tests and range tests (detection of extreme values). The presence and distribution of unknown (non-existent) values in all the variables evaluated were verified. The case registration form was tested to remove ambiguous elements and to protect the data collection instrument. The process was robust and consistent. Statistical analyses were performed using XLSTAT software (New York, NY, USA) for Microsoft Excel version 14.4.0 ((Microsoft Inc., Redmond, WA, USA), and SPSS 20.0 (SPSS Inc®, Chicago, IL, USA).

Continuous quantitative variables are described with the median and interquartile range (IQR). Qualitative variables are described with absolute and relative frequencies (%). To compare the group, in the quantitative variables whose distribution did not show evidence of differing from normal distribution, the Student's *t*-test was used, otherwise the Mann–Whitney U test was used. To compare the percentages, the chi-square test was used for the 2 × 2 contingency tables or, in the case of a low frequency being observed, in some cells of the corresponding table, Fisher's exact test.

Survival analyses were performed using the Kaplan–Meier method and the Cox proportional hazard function.

In-hospital mortality statistics refer to mortality rates at 48 h, patients who were discharged "alive" within 48 h were considered "alive" for the purposes of this analysis. The secondary outcomes were defined as death within 7 days and 30 days of hospital admission. From these estimates, the corresponding area under the receiver operating characteristics curve (AUROC) was obtained for each of the outcomes.

In all the tests carried out, a confidence level of 95% and a value of *p* < 0.05 were considered significant.

## **3. Results**

## *3.1. Patient Baseline*

In total, 1341 cases met the inclusion criteria (out of a total of 3081 patients assessed by EMS) and were part of the cohort analysed (see Figure 1).

**Figure 1.** Flowchart of the participants in the study. <sup>1</sup> In the case of more than one attendance at the emergency department, only the first attendance was analysed. ALS: advanced life support; ED: emergency department.

The median age was 71 years (IQR: 54–83 years), with 38.9% (521 cases) females. The 2 day mortality was 106 patients (7.9%), while it rose to 158 patients (11.8%) at 7 days and 229 patients (17.0%) at 30 days. Regarding the pathologies that the patients included in the study present, it has been observed that the most prevalent diagnosis has been that of cardiovascular origin (29.3%, 393 cases) followed by neurological problems (17.4%, 234 cases), with the ICU admission rate from the ED at 21.3% (285 cases) (see Table 1).


**Table 1.** Demographic, prehospital and hospital clinical outcomes.

Values expressed as the total number (fraction) and medians (25th percentile–75th percentile) as appropriate. Patients included in previous mortality days were also considered for the next period of mortality. The *p* values were calculated with the Mann–Whitney U-test (age, pLA and hLA). The *p* values were calculated with the chi-square test (gender, inpatients, ICU admission and pathology). Other pathology: endocrine, genitourinary, diseases of the blood and the immune system. pLA: prehospital lactate; hLA: hospital lactate; ICU: intensive care unit.

There is a significant correlation between the pLA and hLA levels with two day in hospital mortality. For both values, the median for this mortality range was 5.5 mmol/L (IQR: 4.4–7.6 and 3.3–8.0 mmol/L), while the median in survivors was 3.3 mmol/L (IQR: 2.2–4.8 mmol/L) for pLA and 2.1 mmol/L (IQR: 1.4–3.5 mmol/L) for hLA.

## *3.2. Prognostic Accuracy of pLA and hLA*

The prognostic accuracy of the 2 day mortality of pLA and hLA was good, with an AUROC of 0.800 (95% CI: 0.74–0.85; *p* < 0.001) and 0.819 (95% CI: 0.76–0.86; *p* < 0.001), respectively. Both pLA and hLA lose predictive capacity as time passes. (see Figure 2).

**Figure 2.** Diagnostic performance curves and areas under the curve with 95% confidence intervals for pLA and hLA for 2, 7 and 30 day mortality (in all cases *p* < 0.001). pLA: prehospital lactate; hLA: hospital lactate; AUC: area under the curve; CI: confidence interval.

## *3.3.* Δ*LA and Risk Stratification*

The patients were classified into two groups taking as reference the results of lactate clearance, stratifying the ΔLA in a group with clearance < 10% and another with clearance ≥ 10%. Of all patients, 31.5% (422 cases) had ΔLA with a decrease of <10%, of which a total of 66 patients (15.6%) died. In contrast, in the group with ΔLA ≥ 10%, mortality was only 40 patients (4.4%) (*p* < 0.001) for 2 day mortality (Table 2).


**Table 2.** Correlation between ΔLA and 2 day mortality.

<sup>1</sup> Values expressed as the total number (fraction). ΔLA: delta lactate.

Similarly, and also in line with previous studies, the raw lactate values were segregated based on the initial value and the cut-off point was established at 2 mmol/L to make two comparison groups and measure mortality. With a pLA value < 2 mmol/L, 253 patients (18.9%) were counted, with a single death among them. Altogether, 1088 patients (81.1%) had reference values ≥ 2 mmol/L and in this case the death toll rose to 105 (9.6%). Considering the hLA values with the same cut-off points, 596 (44.4%) and 10 deaths (1.7%) were found with <2 mmol/L. We counted 745 patients (55.6%), of which 96 died (12.9%), with ≥2 mmol/M (see Table 2).

The Kaplan–Meier analysis confirmed significantly longer in-hospital survival at 2 days in patients with lactate ≤ 2 mmol/L compared with patients with higher levels. Survival rates are also consistent with previous results after the analysis at 7 and 30 days. The differences between the survival curves were statistically significant (*p* = 0.001) (Figures 3–5).

**Figure 3.** Kaplan–Meier analysis for 2 day mortality: (**a**) prehospital lactate; (**b**) hospital lactate; and (**c**) lactate clearance. PLAC: prehospital lactate; HLAC: hospital lactate.

**Figure 4.** Kaplan–Meier analysis for 7 day mortality: (**a**) prehospital lactate; (**b**) hospital lactate; and (**c**) lactate clearance. PLAC: prehospital lactate; HLAC: hospital lactate.

**Figure 5.** Kaplan–Meier analysis for 30 day mortality: (**a**) prehospital lactate; (**b**) hospital lactate; and (**c**) lactate clearance. PLAC: prehospital lactate; HLAC: hospital lactate.

## **4. Discussion**

With this study we observed that the measurement of ΔLA can be a quick and easy tool for determining the initial state and the short-term prognosis of a critical patient in an ED. Our results show that both a low lactate level (below 2 mmol/L) and a lactate clearance of more than 10% from the first prehospital determination to the second in the ED is related to an increase in survival.

The concept of lactate clearance was introduced at the end of the last century by Vincent et al. [20] and just as the temporal evolution of lactate and its elimination during resuscitation, this concept has been widely studied in different settings and clinical contexts [21,22].

To our knowledge, this is the first study that analyses the concept of early lactate clearance, in less than one hour, with data collected during prehospital care in the ambulance and in the ED [23,24]. The normalisation of lactate measured in relation to its clearance (ΔLA) was shown to be associated with a lower risk of early death in ED. Poor relative clearance of lactate is an excellent predictor of the risk of early mortality, ahead of the alteration of vital signs [25]. Thus, the measurement of lactate clearance can add useful information for the clinical management of critical patients in an ED.

In line with our results, different authors have observed how a decrease in the lactate level is associated with longer survival (24) and a good response to established treatment [26]. Specifically, Wada et al. and Bhat et al., verified that a decrease in lactate levels among patients attending the ED is associated with longer survival [27,28]; Gotmaker et al. studied the lactate clearance 6 h after the initial determination, obtaining data consistent with ours and asserting that the establishment of this practice can be a very effective tool for assessing the prognosis of critical patients [29]. Something similar is observed by Hguyen and Soliman who analysed this clearance over a longer period such as 12 or 24 h [30,31].

Our study does not only support these previous findings, but also assesses the behaviour of the cohort, with respect to a clearance cut-off point established at 10% of the initial lactate value. This same cut-off has been established by other authors where the elimination of ≥10% lactate at 6, 24 and 48 h is an independent factor related to mortality, even after adjusting for critical status. Lactate clearance is a direct influence factor on survival, more significant than the initial or maximum lactate level reached, in critically ill patients [32,33]. Ladha et al. studied patients admitted to the ICU with a lactate clearance ≥ 10% with respect to the initial value after 6 h, all of whom required less ventilatory support, less need for vasopressor therapy and had a shorter hospital stay [34]. More recent studies showed a higher probability of survival when a second lactate level concentration was less than 3.7 mmol/L, or with a relative lactate clearance ≥ 8% [35].

The first lactate determination in the prehospital setting (pLA) should be complemented with another in-hospital measurement (hLA) upon arrival of the patient. The assessment of ΔLA could help with decision making, reducing the subjectivity of the health worker and complementing the presence of abnormal vital signs [30]. The observation that there has not been a clearance of ≥10% of lactate or the presence of hyperlactacidemia above 2 mmol/L should make us think that perhaps greater intensity should be applied in terms of the resuscitation treatment that we provide. Infected

patients with lactate between 2 and 4 mmol/L have a mortality risk that is twice that of patients with a lactate level less than 2 mmol/L [15]. The early identification of these patients at risk will allow us to improve our response with a reduction in the time of both the necessary diagnostic tests, as well as the establishment of effective treatment [36].

## *Limitations*

Our study has several limitations. Firstly, the study is subject to duration bias, as there was no specific protocol to guide the intervals at which lactate levels are drawn (i.e., time to baseline and time to repeat lactate level). This factor cannot be controlled and could have been delayed for various reasons inherent to the medical activity itself, which would confuse the results with an overestimation of the benefit of screening, although the one hour interval has always been respected. Secondly, the results could be affected by selection bias, since sicker patients with differences in clinical signs may lead to a different response to treatment. However, we did not find significant differences between the elimination group ≥ 10% and the elimination group < 10%, which means that the sample had a severity at the time of lactate extraction. Finally, for future studies, it would be advisable to record in what form and time the treatment is administered, and to explain the differences observed in mortality, perhaps the timing of these interventions is key in the ED.

## **5. Conclusions**

In summary, lactate clearance in the initial moments of ED care appears to be a more reliable prognostic index than a baseline lactate value taken alone. Lactate clearance ≥10% is associated with a lower risk of death in the ED; this value could potentially be used as a guide to determine if a severely injured patient is improving in response to established treatment. Thus, the measurement of lactate clearance appears to be a quick and easy-to-implement tool to determine the initial status and prognosis of the critical patient in an ED. Having the ability to stratify a risk at the earliest stage in critically ill patients can help the ED to more effectively manage the care that these patients need to improve their outcomes.

**Author Contributions:** Conceptualisation, R.L.-I., P.D.B.-I. and F.M.-R.; methodology, P.D.B.-I., J.L.M.-C., A.M.-M.; software, C.D.-F.; formal analysis, M.A.C.V., R.L.-I., P.D.B.-I. and F.M.-R.; investigation, A.M.-M., B.P.-L., C.M.-M., and A.V.; resources, C.M.-M., J.L.M.-C. and B.P.-L.; data curation, M.A.C.V.; writing—original draft preparation, R.L.-I.; writing—review and editing, F.M.-R. visualization, P.D.B.-I.; supervision, J.L.M.-C.; project administration, R.L.-I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Gerencia Regional de Salud de Castilla y León (Spain), grant number GRS 1678/A/18 and GRS 1903/A/19.

**Acknowledgments:** The authors thank the nursing and medical staff from ambulances and the emergency departments of all hospitals, all of them belonging to the public health system of Castilla and León (SACYL), for their assistance in facilitating data collection.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Sponsor's role: none.

## **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Optimization of Case Definitions for Sensitivity as a Preventive Strategy—A Modelling Exemplified with Rapid Diagnostic Test-Based Prevention of Sexual HIV Transmission**

**Andreas Hahn 1, Hagen Frickmann 1,2 and Ulrike Loderstädt 3,\***


**Abstract:** In clinical studies, case definitions are usually designed to optimally match the desired clinical state, because lacking specificity is associated with a risk of bias regarding the study outcome. In preventive medicine, however, high sensitivity is sometimes considered as more critical in order not to overlook infectious individuals, because the latter may be associated with ongoing spread of a transmittable disease. Accordingly, this work was focused on a theoretical model on how the sensitivity of case definitions can be optimized by adding clinical symptoms to diagnostic results for preventive purposes, if the associated reduction in specificity is considered as acceptable. The model was exemplified with an analysis on whether and in how far exposure risk can be reduced by the inclusion of observable symptoms during seroconversion syndrome in case of rapid diagnostic test-based prevention of sexual HIV transmission. The approach provided a high level of safety (negative predictive values close to 1) for the price of a considerably number of false positives (positive predictive values < 0.01 for some subpopulations). When applying such a sensitivityoptimized screening as a "diagnostics as prevention" strategy, the advantages of excellent negative predictive values need to be cautiously balanced against potential undesirable consequences of low positive predictive values.

**Keywords:** rapid diagnostic testing; RDT; sensitivity; modelling; symptoms; transmission prevention; infectious disease; human immunodeficiency virus; HIV

## **1. Introduction**

As recently demonstrated by our group, imperfect accuracy both of diagnostic results [1] and of case definitions [2] can interfere with the outcome of clinical trials in an undesirable way. Accordingly, it is advisable to optimize case definitions for specificity in the most study contexts in order to reduce respective sources of bias [2]. If this is not feasible, sensitivity and specificity of both diagnostic assays [1] and case definitions [2] should at least be known, so diagnostic accuracy-adjusted estimators [3,4] can be applied in order to reduce the effects of associated bias on the study outcomes.

Although optimization of case definitions for specificity may be appropriate for the most instances, however, this does not necessarily apply in all situations. The costs for optimized specificity usually include acceptance of reduced sensitivity [5], implying that a few "cases" may go undetected if highly specific case definitions are applied.

Although both case definitions and diagnostic tests usually try to come as close as possible to the abstract "unknown" truth, "perfect" accuracy for both of them is usually not to be expected in a real-world setting [6]. Because, however, optimization of specificity can usually only be achieved for the price of reduced sensitivity and vice versa [5–7], medical,

**Citation:** Hahn, A.; Frickmann, H.; Loderstädt, U. Optimization of Case Definitions for Sensitivity as a Preventive Strategy—A Modelling Exemplified with Rapid Diagnostic Test-Based Prevention of Sexual HIV Transmission. *Diagnostics* **2021**, *11*, 2079. https://doi.org/10.3390/ diagnostics11112079

Academic Editor: Tivani P. Mashamba-Thompson

Received: 11 October 2021 Accepted: 6 November 2021 Published: 10 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

scientific and even political decision makers will necessarily have to balance potential beneficial and negative consequences of such optimization in the one or the other direction. Abstractly spoken, science may help to quantify the effects of such decisions, but the decision itself within such a balancing will stay a normative one and will depend on the aims of the decision maker.

Optimization for sensitivity rather than for specificity may, e.g., be of interest for public health decision makers in situations when infectious individuals shall not go undetected in order to prevent the further spread of an infectious disease. This is particularly the case, if the consequence for individuals in case of false positive results are mild and can be easily corrected, while severe medical consequences may result from the spread of an infectious disease. Under such circumstances, preventive medical purposes may facilitate balancing more in the direction of optimized sensitivity rather than in the direction of optimized specificity, if the benefits arising from the prevention of the spread of an infectious disease are considered as relevantly more important than potential negative consequences arising from false positive results.

In the exemplarily modelling described here, we introduce how—by themselves—nonspecific observable symptoms may contribute to an increased sensitivity of a case definition which would otherwise rely on a diagnostic assay with imperfect sensitivity alone. Based on the abstract model as presented in the Materials and Methods section, exemplification is conducted with the example of the inclusion of seroconversion syndrome-associated observable symptoms in rapid diagnostic test-based prevention of sexual HIV exposition. Associated advantages and disadvantages are discussed in order to demonstrate potential chances and risks of the abstract concept of sensitivity-optimized case definitions for public health interventions. For this purpose, the HIV pandemic was just exemplarily chosen, because 40 years of experience with the HIV pandemic resulted in the availability of epidemiological details which facilitate modelling approaches based on well-defined epidemiological evidence.

## **2. Materials and Methods**

## *2.1. The Mathematical Background Underlying the Exemplary Modelling*

A sensitive case definition for the identification of an infection may not only include a positive result of a diagnostic test but also a couple of symptoms that have been identified as being associated with this infection. Such a case definition would be fulfilled if the diagnostic test was positive or one or more of the respected symptoms were observed. Based on those assumptions, the overall sensitivity of such case definition is given by:

$$Sensitivity = 1 - \left(1 - Sensitivity\_{Diagnistic\ Test}\right) \times \left(1 - Sensitivity\_{Supports}\right)$$

The specificity is given by:

$$Specificity = Specificity\_{Diagnostic\ Test} \times Specificity\_{Syptopoms}$$

When such a case definition is applied to prevent exposition events towards infections, its positive and negative predictive values (*PPV*, *NPV*) are essential to evaluate its performance. The predictive values *PPV* and *NPV* are given by:

$$PPV = \frac{Sensitivity \times Prevalence}{Sensitivity \times Prevalence + (1 - Specificity) \times (1 - Prevalence)}$$

$$NPV = \frac{Specificity \times (1 - Prevalence)}{Specificity \times (1 - Prevalence) + (1 - Sensitivity) \times Prevalence}$$

With the reciprocal of *PPV* and *NPV*, the number of positive test results needed to get a true positive result and the number of negative test results needed to get a true negative test result are defined, respectively.

With focus on the equation for the *PPV*, it is immediately evident that in case of low prevalence, even perfect sensitivity is not necessarily associated with a good positive predictive value. This is only the case if specificity is almost ideal. Even small deviations from this optimum can lead to a collapse of the positive predictive value. If it is not intended to maximise the sensitivity of the case definition without regard to the positive predictive value, care should be taken in the construction of the case definition to ensure sufficient specificity so that the positive predictive value does not fall below a minimum that is considered as acceptable. The minimum specificity required for a desired sensitivity, prevalence and the minimum positive predictive value still considered as acceptable are given by:

$$Specificity = 1 - \frac{Sensitivity \times Prevalence \times (1 - PPV)}{PPV \times (1 - Prevalence)}$$

The sensitivity and the specificity of the symptoms partly depend on the number of independently distributed symptoms that shall be observed to fulfill the symptoms-related component of the case definition. If there are n symptoms and 1 ≤ *k* ≤ *n* of them have to occur that a patient is "symptomatic" in line with the case definition, then sensitivity and specificity of the symptoms-related component of the case definition are given by:

$$\begin{array}{rcl} \text{Sensitivity} &= P(\ge \text{ } k \text{ } symp \text{ to } occur \mid individual \text{ is infected}) \\ &= 1 - P(< k \text{ } symp \text{ to } occur \mid individual \text{ is infected}) \end{array}$$

$$Specifically = \text{P}($$

It should be noted here that the probability of the occurrence of the symptoms may differ and that they are therefore Poisson binomially distributed.

## *2.2. Assumptions and Prerequisites for the Example of Rapid Diagnostic Test-Based Prevention of Sexual Exposure towards HIV*

2.2.1. Summary of the Testing as Prevention Concept with Focus on HIV

As recently demonstrated by our group based on three previous modelling approaches [8–10] and summarized in a mini-review [11], a combination of self-testing and the testing of potential sexual partners applying traditional or molecular rapid diagnostic testing (RDT) strategies can be a promising approach for the transmission prevention of sexually transmitted infections (STIs) for individuals who do not want to use condoms. As discussed previously [8–12], the effectiveness of such test-based preventive strategies depends on various factors, including the availability of reliable and easy-to-apply diagnostic point-of-care-testing (POCT) solutions even for diagnostic laymen, window-periods of the applied tests as well as the tests' sensitivity and specificity.

In Germany, purchasing of RDTs targeting infections with the human immunodeficiency virus (HIV) by laymen is legally possible since June 2018 [13] as an element of the national strategy for the prevention of HIV transmission. Although self-testing is the intended use of such RDTs, it is nevertheless technically simple to use them for reciprocal testing among potential casual sexual partners who are interested in proving each other "mutually assured" HIV negativity prior to engaging in sexual activity without condom protection. In case of intercourse with sex workers, such condom-free sex is prohibited in Germany since July 2017 by § 32 of the Sex Worker Protection Act ("Prostitutionsschutzgesetz"), demanding condom use in case of all commercial sexual contacts. As sex workers, however, have initially invented the abovementioned "diagnostics as prevention" strategy to protect themselves against HIV transmission in case of agreed unprotected sexual intercourse with their clients long before even the purchasing of HIV RDTs was legally possible [9], it is likely that they will illegally proceed with this strategy in the demimonde. Next to commercial sex work, casual sexual encounters as well among risk groups with high HIV prevalence such as men having sex with men (MSM) may represent situations wherein individuals are potentially interested in reciprocal HIV testing applying RDTs [8,9].

In the first year of implementation of freely availably HIV RDTs in Germany, an estimated quantity of 30,000 tests have been sold and applied [14]. The German society for the support of patients with acquired immunodeficiency syndrome (AIDS) ("Deutsche Aids-Hilfe") considers the strategy of making HIV RDTs freely available for self-testing purposes as a success in the struggle against the ongoing HIV pandemic [15].

Regarding the "diagnostics as prevention" strategy of reciprocal HIV testing by potential sex partners, however, a window-period of traditional immunochromatographic RDTs limits the reliability of this preventive strategy during acute HIV infection, also called seroconversion stage [9]. For this stage of the HIV infection, which is characterized by high viral loads with associated high transmission risk but antibody levels yet below the detection threshold [8,15–20], a combination of molecular RDTs based on polymerase chain reaction (PCR) or loop-mediated isothermal amplification (LAMP) in addition to traditional immunochromatography would be desirable as recently shown to further reduce both the HIV exposition and transmission risk [8,9]. Beyond well-equipped hedonistic clubs, however, availability of molecular HIV testing is presently hardly realistic in the most contexts of risky casual sexual contacts.

2.2.2. Concept of the Inclusion of Seroconversion-Related Symptoms to Increase Sensitivity of the Testing as Prevention Approach

To circumvent the problem of low sensitivity of immunochromatographic RDT-based HIV testing during the seroconversion stage, individuals with affinity to condom-free sex might increase the sensitivity by including symptoms which occur in defined percentages in the course of acute HIV infection/seroconversion syndrome [21,22] in the case definition. Such symptoms could be assessed by direct questioning, but there are reasons which speak against this option. Firstly, medical questioning in a situation of erotic adherence might pose a social challenge. Secondly, as known from strategies trying to avoid reporting bias in studies on sexual medicine [23–27], truthful reports in the context of sexual issues cannot regularly be expected. This could be particularly the case if truthful statements might lead to exclusion from the desired sexual activity. Accordingly, it will be useful to include only symptoms that can be directly checked and verified by the potential sexual partner, before a final decision for or against condom-free sex is made.

As the symptoms of HIV–seroconversion are not specific to acute HIV infections, their inclusion will necessarily lead to a tremendous decrease in specificity of the case definition compared to a case definition based on a positive RDT alone. However, if the consequence of a false positive result is just the use of condoms instead of unprotected sex, prioritizing of sensitivity over specificity may be acceptable in comparison to a slightly higher risk of HIV infection [9]. While in the context of the most studies, optimization for specificity is desirable [1,2], the example provides a situation in which optimization of the case definition for sensitivity seems appropriate. Thereby, the inclusion of directly verifiable disease-associated symptoms into a case definition may help to increase the sensitivity of RDTs, a decision which has to be weighted against lower specificity.

In particular, a case definition for the identification of acute HIV infection (seroconversion syndrome) may include clinical symptoms that occur at an early stage of infection when RDT testing still shows a lack of sensitivity. For the modulation, it has to be assumed that these symptoms are independently distributed. In Table 1, common symptoms of an acute HIV infection with known likelihood of occurrence are given as previously reported [21]. Focussing on symptoms that are sensorially (visibly, tactilely, etc.) verifiable by a third person, such symptom-based case definitions will define an individuum as "positive" in line with the symptom-based case definition if there is at least one of the included symptoms present. Accordingly, the case definition will not be fulfilled if none of the symptoms occurs as chosen for the case definition in Table 1.


**Table 1.** Proportion of common symptoms of acute HIV infection/seroconversion stage according to [21]. Only symptoms sensorially verifiable by a third person were included in the case definition (right column).

P = probability.

The sensitivity of the symptom-related proportion of a case definition based on the included symptoms from Table 1, assuming that at least one of the independently distributed symptoms exists, is given by:

$$Sensitivity = 1 - \prod\_{i=1}^{n} (1 - p\_i)$$

Thereby, *pi* is the likelihood of symptom occurrence in the course of acute HIV infection. Assuming the likelihoods in Table 1 for the included symptoms, the sensitivity of this case definition is 0.94.

The specificity of this case definition strictly depends on the distribution of the symptoms among the non-infected population. Thereby, "non-infected" means that an individuum is not in an acute stage of HIV infection. Accordingly, the specificity of the case definition is the likelihood that none of the included symptoms from Table 1 occurs in the non-infected population and is given by:

$$Specificity = \prod\_{i=1}^{n} (1 - p\_i)$$

Thereby, (1 − *pi*) is the likelihood that a symptom *i* will not occur within the noninfected population.

Since there are no reliable information on the distribution of the most of those moreor-less non-specific symptoms among the non-infected population, the model was adapted to the following different assumptions of symptom distribution among the non-infected individuals: Assuming possible likelihoods that at least one of the chosen symptoms occurs

in a non-infected individual are given by 0.01%, 0.1%, 1%, and 10%, the resulting specificity of the case definition is given by 0.9999, 0.999, 0.99, and 0.9, respectively.

When a case definition as given above is applied to prevent sexually transmitted HIV infections, its positive and negative predictive values (*PPV*, *NPV*) are essential to evaluate its performance.

2.2.3. Assumption Regarding Prevalence and Incidence of HIV Infections as Well as Description of the Stages of HIV Infection as Applied for the Modelling

For Germany, prevalence and incidence of HIV infection for the exemplarily chosen pre-pandemic year 2015 (without differentiation between acute and non-acute HIV infection) are given in Table 2.

**Table 2.** Prevalence and incidence of HIV in Germany in 2015 as described by the Robert Koch Institute, i.e., the central institution of the German federal government with responsibility for disease monitoring and prevention.


HET = heterosexual. MSM = men having sex with men.

The frequency of acute HIV infections can be estimated based of the cumulative duration of each stage of the HIV infection as given in Table 3.


**Table 3.** Viral load by stage of infection according to [18–20] as summarized by our group in [8].

Based on a lack of sensitivity of the RDT chosen for the modelling [9] in the first month of a HIV infection, the incidence of HIV was weighted by the factor 31/365. This assumption results in a weighted incidence of 31.0 females, 31.9 heterosexual males, and 186.9 men who have sex with men (MSM) for the year 2015.

For the diagnostic performance of the assessed Ab/Ag RDT [9,12], the following diagnostic sensitivity and diagnostic specificity after day 10 for the antigen component of the RDT and after day 31 for the antibody component of the RDT were assumed as described elsewhere [9]:


For the combination of the symptom component of the case definition *S* and the RDT component of the case definition, positive and negative predictive values *PPVC* and *NPVC* are given by:

$$PPV\_c = \frac{(Se\_S + Se\_{RDT} - Se\_S \times Se\_{RDT}) \times Precv\_{HIV}}{S + RDT - S \times RDT}$$

$$NPV\_C = \frac{NPV\_S \times NPV\_{RDT} \times (1 - S) \times (1 - RDT)}{(1 - S) \times (1 - RDT)} = NPV\_S \times NPV\_{RDT}$$

Thereby, *S* and *RDT* indicate the expected value of a positive result of the symptombased case definition or the RDT-based case definition, respectively.

Further, it is assumed that the distribution of the symptoms representing the symptombased case definition among HIV infected individuals after day 31 of infection is comparable to the distribution of seroconversion-like symptoms among non-HIV-infected individuals, because the seroconversion stage is close to its end or over. In addition, it is taken for granted that the symptom-based case definition and the RDT-based case definition are stochastically independent.

## **3. Results**

## *Exemplary Modelling of a Sensitivity-Optimized Case Definition Combining Rapid Diagnostic Test Results with Seroconversion-Associated Symptoms for the Prevention of Sexual HIV Exposition*

Based on the assumptions above, the symptoms component of the case definition for the identification of an acute HIV infection results in very low positive predictive values for females and heterosexual males. Its application in the MSM community alone is associated with acceptable positive predictive values if the prevalence rate of the symptoms defining the case definition is very low among the non-infected individuals. The latter means that the prevalence rate of occurrence of at least one of the symptoms in the non-infected population is 0.001 or lower. In this case, the likelihood that a positive result is correct can be expected to be 0.6843. In other words, 1.47 individuals have to fulfill this element of the case-definition in this situation to get one correctly positive test result. In females and heterosexual males, the positive predictive value of such a case definition is almost zero. On the other hand, the negative predictive values are equal to one over all populations and thus, they are also identical with the pretest probability (Table 4).


**Table 4.** Positive and negative predictive values of the symptoms component of the case definition depending on various assumed specificity rates until day 31.

HET = heterosexual males. MSM = men having sex with men.

Additionally, for stages of HIV infection after day 10, the positive predictive values for Ag/Ab RDT-based case definitions are very low in the female and heterosexual male population but much higher than in the scenario for newly infected individuals. Especially for the MSM population, positive and negative predictive values of the RDT-based approach

are very high (Table 5). As shown for the symptom-based approach above, the negative predictive values are high over all populations.



HET = heterosexual males. MSM = men having sex with men.

Although the sensitivity of the symptom-related component of the case definition is 0.94 until day 31, the weighted sensitivity over all stages of HIV infection reduces it to 0.2% for females and heterosexual males and to 0.3% for the MSM population if it is interpreted as a diagnostic test for HIV infection in general and if it is assumed that the distribution of seroconversion-like symptoms of this case definition among the HIV infected population after day 31 is the same as among non-infected individuals. The diagnostic sensitivity of the antigen-component of the RDT (Ag) as a diagnostic test weighted over all stages of HIV infection is reduced to 0 while its specificity is increased to one. For the antibody-RDT-component (Ab), sensitivity is slightly reduced while the specificity is slightly increased (sensitivity and specificity of 0.971 and 0.998 in females and heterosexual males, respectively, as well as 0.970 and 0.999 in the MSM population, respectively).

Combining the symptom-related component of the case definition with the RDT component of the case definition as a diagnostic test for HIV in general provides higher positive predictive values than separately assessed elements of the case definition but remains at a very low level among females and heterosexual males. Among the MSM population, the combined case definitions result in appropriate positive predictive values if the symptom distribution among the non-infected individuals is up to 1% or lower.

As the latter distribution of the symptoms of the symptom-based case definition is uncertain, the Ab/Ag-RDT can be proposed as the most reliable test strategy among the MSM population (Table 6).

**Table 6.** Positive and negative predictive values for a combined case definition including both symptoms and Ag/Ab RDT results.


HET = heterosexual males. MSM = men having sex with men.

## **4. Discussion**

The modelling-based study presented here had a number of aims. Firstly, a model was designed for the increase in sensitivity of case definitions by compensating for the limited sensitivity of a diagnostic test in the early stage of a disease by the inclusion of known symptoms of the respective disease stage. The idea was that such a model might be useful for RDT-based exposition prevention in a pandemic, a concept which has been widely used for the management of the SARS-CoV-2 (severe acute respiratory syndrome-coronavirus 2) pandemic [28] and mostly applying rapid-diagnostic tests with imperfect diagnostic accuracy [29–31]. Accordingly, case definitions in our modelling were not optimized for specificity [1,2], as it is usually desirable in case of clinical trials, but for sensitivity, as the sole aim was the reduction in the exposition risk.

Secondly, the model was tested with a specific example. Due to longer experience with the respective pandemic and thus higher reliability of available datasets for the modelling, the model was not exemplified with the SARS-CoV-2 pandemic but with the HIV-pandemic [32].

As expected, optimization of the case definition for sensitivity had both beneficial and undesirable effects. Based on the known likelihood of defined objectifiable symptoms of HIV seroconversion syndrome [21,22] and the test characteristics of a common HIV RDT targeting both gp24 antigen and HIV-specific antibodies as extracted from a metaanalysis [12], an increase in sensitivity of the case definition "potential HIV seroconversion syndrome" was observed from 12% in case of sole reliance on the RDT results to 94% if objectifiable and verifiable symptoms were included. Thereby, it is of course debatable whether or not the included symptoms are really "recognizable" for medical laymen without respective diagnostic experience, so the practical effect will most likely be lower than the hypothetical one.

Lacking reliable data on the common distribution of the included non-disease-specific, usually mild symptoms in the non-HIV-infected population made an assessment of the specificity of the combined case definition impossible, so only assumptions could be made. Due to the lacking disease-specificity of the included symptoms, however, it has to be assumed that the specificity of such a case definition will be very low, which is a major and expected disadvantage of the approach.

More than this, when applied, for example, to the German "standard" population with an extremely low number of incidental people in the very early stages of HIV infection, even the uncertainty regarding the exact specificity value is practically hardly relevant for the resulting predictive values: The negative predictive value is virtually always close to 1, the positive predictive value is always virtually 0 due to the extremely low number of infected people in the early phase by applying time-weighted incidence for the calculations. Accordingly, the practical information gained when using such a combined case definition, i.e., its reliability for the diagnosis of the HIV seroconversion syndrome, is practical zero with focus on both the positive as well as the negative predictive value.

This, however, does not apply to the exposure risk. In spite of poor predictability of HIV seroconversion syndrome, the exposure probability could be reduced if the case definition was applied correctly. Whereby, however, it would be accepted that the proportion of false positives was of course enormous. If the consequence of this relevant limitation is just a switch from non-protected to protected sexual intercourse; however, the extremely low positive predictive value may be considered acceptable for potential sexual partners willing to protect themselves by reciprocal testing.

With focus on the quantitative dimension of risk reduction in case of the HIV– seroconversion example, it can be concluded that the case definition amended by nonspecific symptoms does not offer a relevant increase in safety, as the initial pretest probability of HIV seroconversion is simply too low within the average German population. For high-risk populations including men having sex with men, the risk reduction is slightly better, making such an approach with an extremely high sensitivity potentially useful. So, in case of doubt in high-risk communities, a respective high-sensitivity-case-definition

might be considered. However, the effect on exposure risk reduction in addition to RDT testing alone will be within the homeopathic range, in particular in case of heterosexual contacts, and only slightly better in the MSM setting.

As exemplified with the HIV pandemic, optimization of case definitions for sensitivity by adding non-specific clinical symptoms [21,22] even to highly specific diagnostic tests [12] can have deleterious consequences on the predictive values. This particularly applies in case of low prevalence of the assessed medical condition and, accordingly, a resulting low pre-test probability. The professional decision on whether or not such an approach may nevertheless be acceptable in a pandemic will largely depend on the expected medical consequences in case of a transmission event.

In the abovementioned example, the still considerable medical consequences of acquiring an HIV infection in terms of requirement for lifelong medical treatment may be balanced against the minor inconvenience of switching from non-protected to protected sexual intercourse. So, the consequences of the high likelihood of false positive results may be considered as acceptable by individuals applying such a sensitivity-optimized "diagnostics as prevention"-based approach of reciprocal RDT-based HIV-testing.

If, however, medical or social consequences of a false positive result are more severe, e.g., defining a need for long isolation periods or quarantine periods for contact persons in a pandemic caused by pathogens other than HIV, the ethical balancing will become more complex.

## **5. Conclusions**

As demonstrated by the model and the example, sensitivity of RDT-based diagnosis in pandemic situations can be considerably increased if non-specific clinical symptoms are included. In particular in case of low prevalence of the diagnosed infectious disease and thus poor pre-test probability, however, the predictive values can be tremendously deteriorated, but the exposure prevention effect can still be increased. Thereby, it has to be decided—balancing both the medical consequences of a transmission event and the social consequences of a false positive result—whether the associated high probability of false positive results in case of applying such case definitions appears justified or not.

The presented modelling has a number of implications for public health decisions in the course of a pandemic. The inclusion of symptoms in case definitions is of particular interest when the former can be clearly identified without the need for medically trained personnel, so that simple mass application seems realistic. The use of such case definitions including symptoms appears to be particularly useful when infectivity is already present before diagnostic detectability, for example by means of rapid tests, because symptoms have already developed to some extent within the diagnostic window period, as illustrated by the example of HIV. However, if the occurrence of symptoms and the detectability of the disease by test assays with poor assay specificity coincide, it seems reasonable to change the linkage of these two components of the case definition from "or" to "and", because an increase in specificity is then advisable. The verification of the suitability of such a case definition optimized with respect to specificity instead of sensitivity is still pending and should be investigated in future studies.

**Author Contributions:** Conceptualization, A.H., H.F. and U.L.; methodology, A.H.; software, A.H.; validation, A.H.; formal analysis, A.H.; investigation, A.H.; resources, A.H. and U.L.; data curation, A.H.; writing—original draft preparation, A.H., H.F. and U.L.; writing—review and editing, A.H., H.F. and U.L.; visualization, A.H.; project administration, H.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** We acknowledge support by the Open Access Publication Funds of the University of Göttingen.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** All relevant data are provided in the manuscript and its tables.

**Conflicts of Interest:** The sponsors did not have any role in the collection, analysis, or interpretation of data, in the writing of the report, or in the decision to submit the article for publication.

## **References**


## *Review* **Systematic Review and Meta-Analysis of the Diagnostic Accuracy of Mobile-Linked Point-of-Care Diagnostics in Sub-Saharan Africa**

**Ernest Osei 1,\*, Sphamandla Josias Nkambule 1, Portia Nelisiwe Vezi <sup>1</sup> and Tivani P. Mashamba-Thompson 1,2**


**\*** Correspondence: 218086551@stu.ukzn.ac.za or ernestosei56@gmail.com; Tel.: +233-242-012-953

**Abstract:** Mobile health devices are emerging applications that could help deliver point-of-care (POC) diagnosis, particularly in settings with limited laboratory infrastructure, such as Sub-Saharan Africa (SSA). The advent of Severe acute respiratory syndrome coronavirus 2 has resulted in an increased deployment and use of mHealth-linked POC diagnostics in SSA. We performed a systematic review and meta-analysis to evaluate the accuracy of mobile-linked point-of-care diagnostics in SSA. Our systematic review and meta-analysis were guided by the Preferred Reporting Items requirements for Systematic Reviews and Meta-Analysis. We exhaustively searched PubMed, Science Direct, Google Scholar, MEDLINE, and CINAHL with full text via EBSCOhost databases, from mHealth inception to March 2021. The statistical analyses were conducted using OpenMeta-Analyst software. All 11 included studies were considered for the meta-analysis. The included studies focused on malaria infections, *Schistosoma haematobium*, *Schistosoma mansoni*, soil-transmitted helminths, and *Trichuris trichiura*. The pooled summary of sensitivity and specificity estimates were moderate compared to those of the reference representing the gold standard. The overall pooled estimates of sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio of mobile-linked POC diagnostic devices were as follows: 0.499 (95% CI: 0.458–0.541), 0.535 (95% CI: 0.401–0.663), 0.952 (95% CI: 0.60–1.324), 1.381 (95% CI: 0.391–4.879), and 0.944 (95% CI: 0.579–1.538), respectively. Evidence shows that the diagnostic accuracy of mobile-linked POC diagnostics in detecting infections in SSA is presently moderate. Future research is recommended to evaluate mHealth devices' diagnostic potential using devices with excellent sensitivities and specificities for diagnosing diseases in this setting.

**Keywords:** mHealth devices; diagnosis; accuracy; sensitivity; specificity; sub-Saharan Africa

## **1. Introduction**

Currently, Sub-Saharan Africa (SSA) bears the highest disease burden worldwide [1]. The high rate of infectious diseases, high recurrence of epidemics, increasing growth of chronic diseases, weak healthcare systems, insufficient funds to support healthcare, limited skilled health professionals, and poor healthcare infrastructure pose a significant challenge in improving healthcare provision in SSA [2–4]. Most patients have limited or no access to healthcare clinics and even essential healthcare services [2]. With these challenges, digital health such as mobile health (mHealth) applications have demonstrated their potentials in screening communicable and non-communicable diseases at point-of-care diagnostics globally, including SSA [5–8]. mHealth technology is considered one of the emerging diagnostic tools or recognized as an enabling technology for disease diagnosis [1,9,10]. In this study, we define mHealth as the use of mobile health devices such as smartphones, tablets, and others as diagnostic tools to diagnose existing disease conditions in patients [11].

**Citation:** Osei, E.; Nkambule, S.J.; Vezi, P.N.; Mashamba-Thompson, T.P. Systematic Review and Meta-Analysis of the Diagnostic Accuracy of Mobile-Linked Point-of-Care Diagnostics in Sub-Saharan Africa. *Diagnostics* **2021**, *11*, 1081. https://doi.org/10.3390/ diagnostics11061081

Academic Editor: Chao-Min Cheng

Received: 25 May 2021 Accepted: 8 June 2021 Published: 12 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

The current global outbreak of the novel Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections has overstretched many healthcare systems, and its implications are still unfolding. With the considerably increasing number of cases and limited available resources, there is a growing need for deployment of scalable solutions such as digital health technologies, including mHealth applications, to monitor and manage the pandemic [5,9]. A recent study in the USA showed that mHealth applications were used to screen healthcare workers for SARS-CoV-2 symptoms to control the spread of the infection [9]. Other studies conducted in the USA, Canada, and Taiwan have also demonstrated the use mHealth for preliminary screening and early detection of possible SARS-CoV-2-infected persons and accelerating linkage to care [10,12,13].

We defined disease diagnosis as the process of identifying a health condition, disorder, or problem by a systematic analysis of a patient's background or history, examining the symptoms, evaluating the test results, and investigating the probable causes [14]. The diagnosis of disease conditions can be performed accurately or inaccurately by health professionals, patients, and other recognized groups. In this study, diagnostic accuracy can generally be defined as the actual results that contain both true positives (sensitivity) and true negatives (specificity) of a disease condition in a population [15]. Diagnostic accuracy can further be described as a test's ability to discriminate between the target disease condition and health [16].

In low- and middle-income countries (LMICs), several mobile health techniques are being utilized to support healthcare delivery. Studies in SSA revealed that mobile health techniques such as short message service (SMS), voice/phone calls, and mobile apps are predominantly employed to support healthcare delivery [3,11,17]. For instance, recently, mobile phone devices are used to capture images that are processed immediately and analyzed using smart algorithms for disease diagnosis [6,7]. In Botswana, mobile phones are used for diagnostics accuracy of photographs of plain film test X-rays digitally [7]. In SSA, healthcare professionals employed the SMS technique for educating and creating awareness on treatment methods, management of diseases, and availability of health services [8]. Similarly, SMS and voice calls are used to remotely monitor chronic conditions, communicate, and train healthcare professionals, track pandemic and epidemic outbreaks, and data collection [8,11]. Additionally, in SSA and other settings, mobile health techniques such as mobile apps allow the community healthcare workers to enter patients' symptoms into the app, diagnose illness, and give treatment recommendations [2,4,8]. Furthermore, research has demonstrated that mHealth applications like mobile apps are primarily used for collecting clinical data of patients and healthcare systems to assist in formulating health policies [8,18]. Studies have demonstrated that the short message service technique is the most used mHealth application to support healthcare delivery in SSA [19–21]. The evidence available shows that most of these mHealth techniques are based on optical detection methods [6,7].

Our scoping review aimed at mapping evidence on mHealth applications to diagnose diseases and support treatment procedures by healthcare workers in SSA [22]. The results showed that mHealth applications are available and are being used to support healthcare services by health professionals. The results demonstrated that mHealth applications are being used for diagnosing certain disease conditions in SSA. The results further indicated that mHealth applications are being utilized to manage HIV, TB, cancer, and hypertension cases in SSA [22]. In recent times, mobile health devices have been employed to provide accurate and rapid diagnosis of diseases at POC diagnostics, which is critical to provide effective and life-saving treatments [23–26]. Other studies have also demonstrated that access to a simple mHealth device at POC diagnostics can potentially transform individuals' health behavior and improve people's preventive interventions in hard-to-reach communities [27,28]. Similar studies revealed that mHealth devices had been used in resource-poor settings at POC diagnostics to detect recent infectious Ebola, Severe Acute Respiratory Syndrome (SARS), and Zika viruses to help in the early treatment of such cases [29–32]. Although the advent of mobile-linked diagnostics at point-of-care in resource-limited settings helps improve access to healthcare and reduce healthcare inequalities [23,24], there

is limited evidence on their diagnostic accuracy. Therefore, we performed this systematic review and meta-analysis to evaluate mobile-linked POC diagnostics' accuracy in SSA.

## **2. Materials and Methods**

The review followed the Preferred Reporting Items requirements for Systematic Reviews and Meta-Analysis (PRISMA) [33]. The Population, Intervention, Comparison, and Outcome (PICO) framework for determining the primary research question eligibility (Table 1) was followed.

**Table 1.** PICO framework for determining the eligibility of the research question.


The primary research question was: What is the evidence on the diagnostic accuracy of mobile-linked POC diagnostics in Sub-Saharan Africa?

## *2.1. Search Strategy*

An electronic search was carried out to identify all relevant published descriptive quantitative studies, randomized controlled trials, non-randomized controlled trials, and mixed-method studies to answer the review question. As part of our search criteria, database searches were conducted from mHealth technology inception to July 2019. They were updated in March 2021 using PubMed, Science Direct, Google Scholar, MEDLINE, and CINAHL with full text via EBSCOhost databases. Reference lists of all included studies eligible for inclusion were also searched for relevant potential articles. Boolean terms (AND, OR) and MeSH (Medical Subject Headings) terms which formed part of the search strategy were used. The keywords used for the search included: "mHealth apps", "mHealth devices", "diagnostic", "accuracy", "sensitivity", "specificity", "health workers" and "sub-Saharan Africa" (Supplementary file S1). During the search, limitations such as date and language were removed.

## *2.2. Study Selection*

Following databases search for all the relevant studies, the principal investigator (EO) initially screened all titles of articles identified via the search strategy. All the eligible study titles were then exported to an Endnote X9 library specifically designed for this review. All duplicates identified were deleted, and the Endnote library was shared with the review team for abstract screening, which E.O. and P.N.V. performed in parallel. All discrepancies between the reviewers' results following abstract screening were resolved through discussion until consensus was reached. Included studies following abstract screening were included for full-article screening performed by two reviewers, E.O. and P.N.V., independently. T.PM.-T., a third reviewer, was invited to resolve all the discrepancies in screeners' results following the full-text screening. The screening was guided by the eligibility criteria presented below:

## *2.3. Eligibility Criteria*

To ensure that all relevant evidence sources were identified and selected for our review, the study selection process was guided by the eligibility criteria specified under the inclusion and exclusion criteria.

## 2.3.1. Inclusion Criteria

The following criteria were used:


## 2.3.2. Exclusion Criteria

The following were excluded:


## *2.4. Data Extraction*

We designed a data extraction tool specifically for this review to extract all the relevant data from the included primary studies. The data for the analysis extracted from the included primary studies were organized in two sections: basic information and the primary study outcomes. The first section had the name of the author(s), date of publication, the aim of the study, country of research, study design, geographical settings, study setting, study population, sample size, type of mobile-linked POC diagnostics, key findings and conclusions. The second section also included true-positive values, false-positive values, true-negative values, false-negative values, sensitivity, specificity from each of the included primary studies, and a 2 × 2 table was constructed. E.O. and T.P.M.-T. independently conducted the included studies' data extraction using the designed standard data extraction tool. A discussion resolved discrepancies between the reviewers' responses until a consensus was reached.

## *2.5. Assessment of Methodological Quality*

The Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool was employed to assess the quality of all the included primary studies [34]. Quadas-2 is a wellstructured tool recommended by the Cochrane Collaboration for determining diagnostic accuracy studies by evaluating them in four main domains: patient selection, index test, reference standard, and flow and timing [34]. The included primary studies' risk of bias was comprehensively assessed independently by two reviewers (E.O. and T.P.M.-T.). All the disagreements in their assessment were resolved via a discussion.

## *2.6. Data Analysis*

The meta-analysis of diagnostic accuracy was considered for studies whose sensitivity and specificity had been evaluated. Statistical analyses were all performed using the Rbased software Open Meta-Analyst [35]. A random-effects model (DerSimonian-Laird) was used to calculate the pooled sensitivity, specificity, and diagnostic odds ratio (DOR) with a 95% confidence interval (CI). A summary receiver operating characteristic curve (ROC) was constructed by plotting the individual and summary points of sensitivity and specificity to determine mobile devices' overall diagnostic accuracy. Heterogeneity among the included primary studies was determined using *I* <sup>2</sup> statistics where a score of 25% indicates low, a score of 50% represents moderate, and a score of 75% means high levels of heterogeneity [36]. A *p*-value < 0.05 was employed to demonstrate a statistically significant association in all the analyses.

## **3. Results**

## *3.1. Search*

A total of 29,976 articles were identified from the combined search. Seven hundred fortyeight articles were eligible from the database search. One hundred eight-six duplicates were removed, leaving behind five hundred sixty-two articles suitable for abstract screening. A total of four hundred ninety-nine articles were excluded following the abstract screening. Sixty-three articles were eligible for full-text screening. Fifty-two of them were excluded, as illustrated in Figure 1, showing the PRISMA flow chart of literature search and selection of studies. Finally, 11 articles were included for data extraction and further underwent quantitative meta-analysis.

**Figure 1.** PRISMA flow chart showing literature search and selection of studies.

## *3.2. Characteristics of the Included Articles*

Table 2 illustrates the characteristics of the included studies. A total of 11 articles were reviewed, and all underwent meta-analysis. Three of the included articles were conducted in Côte d'Ivoire [37,38], two in Ghana [39,40], two in Uganda [41,42], two in Sudan [43,44], one in Tanzania [45], and one in Ethiopia [46]. Sample sizes ranged from 50 to 1530 persons. Out of 11 studies, only 1 was a cohort study, and 10 were cross-sectional studies. All the included primary studies presented findings on the diagnostic accuracy of mobile-linked POC diagnostics in SSA. In terms of geographical settings, eight of the included studies were conducted in rural locations [37–41,45,46], while three were conducted in urban settings [42–44]. All the 11 included studies were conducted in English language from 2010 to 2017.


**Table 2.**

Characteristics

 of the included studies.


**Table 2.** *Cont.*

## *3.3. Assessment of Risk and Applicability*

Table 3 shows the risk of bias and applicability concern assessment of the included studies using the QUADAS-2 tool. The results illustrate a range of findings in the included studies that employed QUADAS-2 as the quality assessment tool [34]. Participants' enrolment in all the included studies was not based on random sampling or consecutive techniques regarding the patient selection domain but rather on a convenience approach. Even though it is highly possible that the convenience sampling technique could introduce a high-risk bias, it is unlikely to affect the diagnostic accuracy of mHealth devices. The reference standard domain was found to be at low risk of bias across all the included studies. The index test domain was at low risk of bias for most of the included studies. All the included studies were at low risk of bias in the flow and timing domain. However, all the studies included were at high risk of bias under the patient selection. Concerning the applicability assessment, nine of the included studies were at low risk of bias, while two were found to be a high risk of bias. Figure 2 displays the graphical results of the included studies from the QUADAS-2 assessment tool.

**Table 3.** Summary of methodological quality assessed with the QUADAS-2.



## *3.4. Diagnostic Accuracy of Mobile-Linked Diagnostic Devices*

Table 4 illustrates true-positive, false-negative, false-positive, true-negative results and their corresponding sensitivity and specificity values for mobile-linked POC diagnostic devices for detecting disease conditions. The summary estimates of sensitivity and specificity of mobile-linked devices were 0.499 (95% CI: 0.458–0.541) and 0.535 (95% CI: 0.401–0.663), respectively (Figure 3A,B). The pooled estimates of specificity and sensitivity were statistically significant at the meta-analysis level. The individual pooled and summary estimates of sensitivity and specificity at the 95% CI region for all the included studies of mobilelinked POC diagnostic devices are presented in an ROC graph (Figure 4). The overall pooled estimates of the positive likelihood ratio (PLR) and negative likelihood ratio (NLR) were 0.952 (95%CI: 0.60–1.324) and 1.381 (95%CI: 0.391–4.879), respectively (Figure 5). Heterogeneity was determined as statistically insignificant, as I2 = 35.6% (*p* = 0.098) for the degree of inconsistency. The ROC curve analysis demonstrated a significantly moderate diagnostic performance of the mobile-linked POC diagnostic devices. The diagnostic odds ratio (DOR) for mobile-linked POC diagnostic devices' accuracy was found to be OR = 0.944 (95% CI: 0.579–1.538) (Figure 6). Hence, the overall effect estimate of the study at the meta-analysis level was statistically insignificant.


**Figure 3.** (**A**) Forest plots of pooled sensitivity and specificity estimates for all included studies of mobile-linked diagnostic devices; (**B**) Forest plots of pooled specificity estimates for all included studies of mobile-linked diagnostic devices.

**Figure 4.** ROC graph of the included studies of mobile-linked POC diagnostic devices.

**Figure 6.** Diagnostic odds ratio forest plot of the included studies of mobile-linked diagnostic devices.

## **4. Discussion**

The evidence available from this study showed a moderate diagnostic accuracy of mobile-linked POC diagnostics in Sub-Saharan Africa. This systematic review's objective was to evaluate the diagnostic accuracy of mobile-linked POC diagnostics in SSA. We found that mobile-linked POC diagnostics' overall sensitivity for disease detections was 49.9%, and specificity was 53.5%. The meta-analysis results indicated a moderate diagnostic accuracy of mobile-linked POC diagnostic for disease detections in SSA. The ROC curve also confirmed the average diagnostic performance of these mobile-linked POC diagnostic devices. This means that mobile-linked POC diagnostics have less sensitivity and specificity abilities than the cut-off value of the gold standard described by the World Health Organization (WHO) [47]. We performed a sub-group analysis of the included studies to determine the rate of sensitivities and specificities of similar disease outcomes. A cursory examination of seven included studies that used mobile-linked POC diagnostic devices to detect malaria infections found moderate sensitivity and specificity estimates of 0.500 (95% CI: 0.352–0.648) and 0.500 (95% CI: 0.019–0.981) compared to the cut-off value of the gold standard light microscope described as an effective diagnostic tool [47].

The results also demonstrated that two studies that used mobile-linked POC diagnostic devices to detect *Schistosoma mansoni* found an average sensitivity estimate of 0.500 (95% CI: 0.380–0.620) and a low specificity estimate of 0.010 (95% CI: 0.001–0.136) compared to the gold standard conventional light microscope [47]. Again, the results illustrated that mobile-linked POC diagnostic devices for detecting *Schistosoma haematobium* infections found a low sensitivity estimate of 0.008 (95% CI: 0.409–0.601) and an average specificity estimate of 0.500 (95% CI: 0.019–0.981) compared to the gold standard conventional light microscope [47]. Additionally, the results indicated that two studies that used mobile-linked POC diagnostic devices to diagnose *Trichuris trichiura* infections found moderate sensitivity and specificity estimates of 0.511 (95% CI: 0.429–0.592) and 0.500 (95% CI: 0.388–0.612) compared to the gold standard light microscope [47]. These mobilelinked POC diagnostic devices providing moderate sensitivity and specificity estimates proved that such devices are below the cut-off point compared to the gold standard light microscope. The moderate diagnostic abilities of mobile-linked POC diagnostic devices for infectious and non-infectious diseases could also be attributed to the first-generation mobile phone microscopes employed in most of the included studies.

A study conducted in some LMICs found the use of mobile phone fluorescence microscopy for detecting waterborne pathogens had an accuracy of 95%, which is not consistent with our study results [48]. Similar studies conducted in Finland and New Zealand illustrated that mobile phone microscopes exhibited high sensitivity for detecting soiltransmitted helminths and *Schistosoma*, which does not agree with our study results [49,50]. Luis Rosado et al. carried out another study in Portugal where s mobile phone microscope displayed higher sensitivity and specificity for diagnosing malaria infections, at variance with this study's results [51]. A survey conducted in the USA by Paul Slusarewicz et al. revealed that mobile phone microscopes detected parasite eggs in mammalian feces with high sensitivity and specificity, which disagrees with this study's findings [52]. A study conducted in Sweden revealed that mobile phone microscopes could be used extensively for clinical diagnostics when their sensitivities reach or exceed the 80% threshold [49]. Studies conducted in the USA have demonstrated that mobile handheld devices had a high diagnostic accuracy at POC diagnostics for detecting coronary stenosis and other disease conditions [26,53].

This review study included studies carried out in different geographical settings, given an exhaustive overview of the diagnostic accuracy of mobile-linked POC diagnostic devices in SSA. Date and language limitations were removed from this review study to capture all the essential literature on mobile-linked POC diagnostic devices' diagnostic accuracy in SSA. Nonetheless, a piece of evidence on mobile-linked POC diagnostic devices' diagnostic accuracy in SSA might have existed under different contexts that were not included in the study. This review was limited to studies that used quantitative methods, since this study focused on the diagnostic accuracy of mobile-linked POC diagnostic devices in SSA. The systematic review was also limited to studies conducted in SSA and could not be made to represent the entire world.

The results illustrated that most of the studies were conducted in rural settings where there is no access or little access to standard laboratory facilities. This will benefit such rural inhabitants by improving their health conditions if these activities are often conducted in such areas. The study results provided a moderate diagnostic yield of disease conditions and may not encourage healthcare professionals to rely on such devices to support healthcare provision continually. This means that more technologically advanced mobile-linked POC diagnostic devices, well validated with excellent sensitivities and specificities, should be made available to these healthcare professionals and other users.

The results suggested that most of the studies that used first-generation mobile phones attached to microscopes provided a modest diagnostic yield of infectious and non-infectious diseases in resource-poor settings. We recommend future research on using low-cost technologically advanced mobile phone microscopes at POC in resource-constrained settings that may improve their diagnostic capabilities. The results also indicated that mobile-linked POC diagnostic devices' diagnostic accuracy in detecting infectious and non-infectious diseases was found only in six SSA countries. We, therefore, encourage more countries in SSA to employ these mobile-linked POC diagnostic devices to assist in diagnosing more infectious and non-infectious diseases, especially in remote areas.

The QUADAS-2 results showed a high risk of bias under the patient selection domain, which means that patients were selected not based on all consecutive or random sampling techniques. Employing any of these techniques means that eligible patients with suspected disease conditions were more likely to be chosen than those without any condition. In the included studies, inappropriate exclusions were made, which could have led to overoptimistic estimates of diagnostic accuracy. Studies that used consecutive patients with confirmed diagnoses were more likely to show greater sensitivity than those that included patients with suspected conditions. The low risk of bias under the index test domain for most of the included studies was because the index test results were interpreted without knowing the reference standard results. The low risk of bias under the reference standard domain means that the estimates of test accuracy were based on the reference standard with 100% sensitivity and specificity. It also means that the reference standard results were interpreted without the knowledge of the test index results. The low risk of bias in the flow and timing domain means that a reasonable time interval between index test and reference standard was given. This helped to determine the presence or absence of a target condition in the included studies. In cases where there is a bit of delay between the index test and reference standard, a possible misclassification of a disease condition may occur due to either recovery or deterioration of such condition.

## **5. Conclusions**

Mobile-linked POC diagnostic devices can improve healthcare provision quality in clinical care to diagnose diseases in resource-constrained SSA areas. Current devices have been integrated slowly in routine clinical practice, with innovations such as mobile phone microscopes, machine learning, computer vision, and others that could assist in automatic diagnoses of diseases. The study results illustrated that mobile-linked POC diagnostic devices provided an average diagnostic yield in detecting infectious and non-infectious diseases in SSA. The study results further demonstrated that the first-generation mobile phones employed contributed to the moderate sensitivities and specificities in diagnosing infections in low-resourced SSA settings. Hence, we recommend that much more primary research should be carried out in SSA with mobile-linked POC diagnostic devices. These devices should be technologically advanced and well validated to provide sensitivities and specificities estimates to reach or exceed the 80% threshold. We also recommend that more mHealth diagnostics evaluation studies employ refined mHealth devices with excellent sensitivities and specificities to diagnose existing diseases in SSA.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/diagnostics11061081/s1, Supplementary file S1: Results from the initial database search.

**Author Contributions:** E.O. and T.P.M.-T. conceptualized and designed the study. E.O., P.N.V., and T.P.M.-T. contributed to the abstract, full-article screening, and the included studies' quality assessment. S.J.N. performed the meta-analysis and assisted in the interpretation of the results. E.O. prepared the draft of the study, T.P.M.-T. reviewed the draft critically. E.O. prepared the final draft. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research did not receive any specific grant from funding agencies in public, commercial, or not-for-profit sectors.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors wish to thank the University of KwaZulu-Natal for giving them all the necessary resources in developing this review.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Prospero Registration:** CRD 42020155041.

## **References**


## *Review* **Evidence of TB Services at Primary Healthcare Level during COVID-19: A Scoping Review**

**Thobeka Dlangalala 1,\*, Alfred Musekiwa 1, Alecia Brits 2, Kuhlula Maluleke 1, Ziningi Nobuhle Jaya 1,3, Kabelo Kgarosi <sup>4</sup> and Tivani Mashamba-Thompson <sup>5</sup>**


**Abstract:** Tuberculosis (TB) is still a major public health concern, despite the availability of preventative and curative therapies. Significant progress has been made in the past decade towards its control. However, the emergence of the novel coronavirus disease 2019 (COVID-19) has disrupted numerous essential health services, including those for TB. This scoping review maps the available evidence on TB services at the primary healthcare (PHC) level during the COVID-19 period. A comprehensive literature search was conducted in PubMed, Web of Science, Medline OVID, Medline EBSCO, and Scopus. A total of 820 articles were retrieved from the databases and 21 met the eligibility criteria and were used for data extraction. The emerging themes were the effect of the COVID-19 pandemic on TB services, patient and provider experiences, recommendations for TB services during the COVID-19 period, and the implementation of the recommendations. The review found that the mitigation strategies, as well as fear and stigma experienced at the start of the COVID-19 pandemic may have led to TB cases potentially going undetected, which may threaten TB treatment outcomes. Therefore, efforts must be directed at finding these missing cases and ensuring that PHC facilities are equipped to adequately diagnose and treat them.

**Keywords:** COVID-19; coronavirus; tuberculosis; health services; primary healthcare

## **1. Introduction**

Despite the availability of vaccinations and chemotherapy for prevention and treatment [1], 10 million new cases of tuberculosis (TB) were estimated to have occurred in 2019 [2]. However, only 7.1 million of these cases were found and reported to national TB programmes, leaving a third undetected [3]. In addition, considerably more were not started on an appropriate treatment [1]. These missed cases contribute to the ongoing transmission [4], while prolonged diagnosis and treatment initiation exacerbate disease severity and continued spread [5]. Interrupting transmission through early and accurate detection, rapid treatment initiation, and completion, preferably at the primary healthcare level (PHC), aids efforts in ending the TB epidemic [3,6]. In 2020, COVID-19 emerged, hindering global TB control efforts [7], and sidelining many routine TB services to accommodate the response to the COVID-19 pandemic [8,9]. TB services suffered a sharp decline due to lockdowns. Therefore, limiting access to healthcare and a rise in fear and stigma since the advent of COVID-19 [8,10,11].

**Citation:** Dlangalala, T.; Musekiwa, A.; Brits, A.; Maluleke, K.; Jaya, Z.N.; Kgarosi, K.; Mashamba-Thompson, T. Evidence of TB Services at Primary Healthcare Level during COVID-19: A Scoping Review. *Diagnostics* **2021**, *11*, 2221. https://doi.org/10.3390/ diagnostics11122221

Academic Editor: Xavier Muñoz-Berbel

Received: 20 October 2021 Accepted: 20 November 2021 Published: 27 November 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Studies that predict the potential impact of the COVID-19 pandemic on TB services suggest that temporary disruptions in response to the pandemic will likely affect all aspects of the TB care cascade [12–14]. Even small disruptions to these services could have longterm consequences on TB control [12]. These will especially be felt in high burden countries where TB incidence and mortality have been predicted to increase by 6.3 and 1.4 million, respectively, between 2020 and 2025 [12]. Delays in timely diagnosis and treatment are listed as the potential drivers for these grim outcomes [12,14].

The World Health Organization's (WHO) End TB strategy and the sustainable development goal (SDG) 3.3 aim to end TB through timely diagnosis and treatment, treatment adherence, and preventative therapy [15,16]. The WHO aims to eliminate the TB epidemic by 2035 and has also set short-term milestones to reduce TB deaths and incidence rates by 2020 and 2025 [3,15]. Findings from the TB global health report showed that 2020 milestones were not achieved [3,17]. Similarly, interim targets were set by the United Nations (UN) to diagnose and treat 40 million additional people by 2022 [7]. Although progress towards these goals has been made, it is still below the threshold that would make TB elimination attainable [3,18]. Moreover, it is possible that the small gains made towards controlling TB were disrupted by the COVID-19 pandemic, pushing the global TB targets further into the future [7,19].

As the first point of contact with health services, PHC facilities can reach large proportions of the population. These facilities also promote equitable access to health services and continuity of care and are recognized as a powerful tool for achieving the health SDGs [16,20]. Moreover, the WHO has emphasized that progress towards containing the TB epidemic can accelerate when TB control has been integrated with PHC [21]. Furthermore, high-quality PHC services are an important predictor for whether TB control strategies will realize their promise [22].

Despite the emergence of other public health priorities, such as the COVID-19 pandemic, uninterrupted TB services at the PHC level are crucial for reaching TB targets. Given the novelty of the COVID-19 pandemic, its effects on TB services at the PHC level remain unclear and require further exploration. Therefore, this scoping review mapped evidence on TB services at the PHC level during the COVID-19 pandemic. This evidence will be used to develop the primary research in order to address and improve TB services at the PHC level during the COVID-19 pandemic.

## **2. Materials and Methods**

## *2.1. Overview*

Herein, we conducted a scoping review to map the available evidence on TB services during the COVID-19 era. This scoping review is conducted as part of a larger study that aims to develop a novel approach for improving TB diagnostic services during the pandemic in primary healthcare clinics in high disease burdened settings. A scoping review protocol was registered on the open science framework (OSF) under the title, "Evidence of TB services at primary healthcare level during COVID-19: A scoping review protocol", where it can be accessed via this link: https://osf.io/pq3ba, 15 October 2021. The scoping review was guided by the Arksey and O'Malley framework [23], Levac et al. [24], and the Joanna Briggs Institute 2020 guidelines [25]. The findings of the study were reported according to the Preferred Reporting Items for systematic reviews and meta-analyses extension for scoping reviews (PRISMA-ScR) checklist, Table A1 [26].

Step 1: Identifying the research question

The main research question was: What evidence exists on TB services at the PHC level during the COVID-19 pandemic?

We assessed the eligibility of the research question for a scoping review study by applying the population, concept, and context (PCC) framework, developed by the Joanna Briggs Institute [25], see Table 1.


**Table 1.** PCC framework to determine the eligibility of the research question and guide the selection of studies on TB services during the COVID-19 pandemic.

## Step 2: Identifying relevant studies

We conducted an advanced search using the following five academic databases: PubMed, Web of Science, Medline OVID, Medline EBSCO, and Scopus. Studies were identified using the following keywords and Medical Subject Heading (MeSH) terms: "TB diagnostics", "Health Service" "TB testing" "COVID-19", "SARS-CoV-2", "COVID-19 Pandemic", "COVID-19 era", and "Primary healthcare". A combination of Medical Subject Headings (MeSH) and free word texts of the keywords were used when conducting the searches. WHO and Stop TB partnership websites were accessed for reports and the reference lists of all the included studies were consulted for additional literature. The comprehensive database search was conducted by an experienced librarian to ensure that the best search strategies were used for each database.

Publications that adhere to the following criteria were included:


This review excluded studies based on the following:


Step 3: Selecting studies

The studies were selected in three phases. First, the principal investigator screened the titles of each article using the eligibility criteria as a guide. Eligible articles were exported to an EndNote20 library where duplicates were identified and removed. In the second phase, two independent reviewers screened the abstracts of the included articles using a screening tool based on inclusion and exclusion criteria. The screening tool was piloted and adjusted using 10 articles before the screening process was conducted. The reviewers discussed any discrepancies that arose until they reached a consensus on the articles to select. In the third phase, the two reviewers screened the full texts of the relevant articles using a screening tool guided by the eligibility criteria. Before use, the screening tool was piloted by both screeners, and changes were made accordingly. Discrepancies during full-text screening were resolved by a third reviewer. The level of agreement between the two reviewers was calculated using the Kappa statistic.

Step 4: Charting the data

An electronic data charting form containing variables relevant to the research question was developed. Two independent reviewers then piloted the data extraction tool using 10 of the included studies. The necessary changes were applied according to the feedback given by the reviewers. Data were extracted from the included studies based on the following categories: Author, aim, type of publication, country, type of TB service, and primary healthcare provider.

## *2.2. Quality Appraisal*

We determined the methodological quality of the included studies using the Mixed Methods Appraisal Tool (MMAT) V.2018 software [27]. The particular study design used in each article was appraised, following stipulations by the MMAT guidelines. Once the scores for each study were calculated as a percentage, they were given a specific rank. Studies equal to or below 50% were ranked as low quality, those between 51–75% were deemed average quality, and those ranging from 76–100% were given a high-quality score.

## *2.3. Collating, Summarizing, and Reporting Results*

We employed the thematic analysis to extract relevant evidence to answer our research questions and presented a narrative summary that centered on the emerging themes. The themes that arose most from the included studies were as follows: The consequences of the COVID-19 pandemic on TB services, comparison of TB services before and after the COVID-19 pandemic, patient experiences of TB services during COVID-19, and recommendations for TB services at PHCs during COVID-19.

## **3. Results**

## *3.1. Screening Results*

The selection and exclusion of studies are depicted in the PRISMA-ScR flow chart (Figure 1). Initially, we retrieved 819 articles, 702 from database searches and 117 from Google. Following title screening, we excluded 594 ineligible articles. The 225 remaining articles were imported to Endnote 20. The results retrieved from each database are listed in Table 2. After removing 120 duplicates, 105 articles were eligible for abstract screening. A total of 54 articles were excluded after abstract screening and 51 were eligible for full-text screening. We excluded 30 articles after full-text screening. All of the articles reported findings from the pandemic and articles were excluded if they reported TB services outside of PHC (17), did not mention healthcare setting (9), and combined data on TB services from both PHC and higher healthcare settings (3). In total, 21 articles met the eligibility criteria and were used for data extraction. The responses of the reviewers had a 54.64% agreement versus a 73.77% expected agreement by chance, which equates to a moderate agreement (Kappa statistic = 0.4218, *p*-value < 0.05). The discrepancies from the full-text screening were resolved by a third screener.

## *3.2. Characteristics of the Included Studies*

The characteristics of the included articles are detailed in Table 3. The studies presented evidence on TB services at the PHC level during the COVID-19 era. The findings were conveyed in a variety of formats including letters, editorials, expert opinion, reports, webinars, feature articles, news articles, and traditional research articles. In terms of countries, the included articles were from Portugal [28], Ethiopia [29], Japan [30], China [9], Malawi [31], the United States of America [32], Pakistan [33,34], Nigeria [35–37], India [38–40], South Africa [41–43], one provided recommendations for high burdened settings [44], one presented evidence from LMIC [45], and one study was addressed to all the countries [46]. The primary healthcare settings ranged from clinics, outpatient departments, general practitioner's practices, PHC centers, and pharmacies.

**Figure 1.** Prisma-flow diagram depicting the process of selecting and excluding studies.


## **Table 2.** Results of the database search.

## *3.3. Quality Appraisal*

Only four articles were primary studies presenting empirical evidence and were subject to a methodological quality assessment using the 2018 version of the MMAT tool [27]. The scores ranged from 40–75%. Two studies scored 60% [9,32] and another scored 40% [47] and 70% [35]. Results that scored lower than 51% were considered low quality, 51–75% were of average quality, and high quality if they fell between 76–100%.

## *3.4. Summary of the Evidence*

The themes that emerged from the included studies were, consequences of COVID-19 pandemic on TB services, patient and provider experiences, recommendations and adaptations for TB services during the COVID-19 era, and implementing the recommendations for TB services, respectively.


**Table 3.**

Characteristics

 of the included studies.


**Table 3.** *Cont.*

## 3.4.1. Consequences of the COVID-19 Pandemic on TB Services

Of the 21 included studies, 10 reported on the consequences of the COVID-19 pandemic at various PHC facilities. TB clinics in New York, USA temporarily halted the performance of any new TB tests [32]. A study from a LMIC reported that fewer TB cases were diagnosed due to the difficulty in accessing primary care [45], while a clinic in Nigeria reported that one person came to collect the TB medication during the lockdown [37]. South Africa experienced a 25% drop in access to primary healthcare following the lockdown, as well as a 9% drop in TB testing [41]. Another study in China reported that 75.3% of primary healthcare workers were reallocated from routine services to COVID-19 related work [9]. In a similar manner, clinics from Ethiopia were repurposed as COVID-19 centers [29] or in the case of TB clinics in New York, USA, closed altogether [32]. In Japan, the media reported a shortage of the BCG vaccine in order to claim that it was effective against COVID-19 [30].

A project that brought TB healthcare to the doorstep of a community was abruptly halted after the nationwide lockdown in India [39]. This project was aimed at rendering a neighborhood block TB-free and achieved it by actively finding TB cases and providing point-of-care mobile diagnostic services. The effects were seen by the abrupt drop in TB notifications during the 3 months of the national lockdown. In addition, direct comparisons with the same period from previous years showed a stark contrast. Another study in Nigeria that sought to directly compare TB case notifications and detection rates in the first few months of 2020 compared with the same period from 2019 showed similar results [35]. Another study from Ethiopia showed that patient flow had significantly decreased in the first months of the COVID-19 lockdowns compared with the same period from the previous year [29]. Moreover, TB case notifications at primary healthcare centers in Malawi were shown to be disproportionately lower than at a regional hospital in Malawi [31]. The current evidence shows that the COVID-19 pandemic has created a scenario where fewer TB cases were detected than usual. However, more evidence is required to determine the extent of the potentially missed cases.

## 3.4.2. Patient and Provider Experiences

Four of the included studies recorded the perspectives of healthcare workers and patients. All of the participants struggled to access healthcare facilities. Rumors on the closure of certain facilities meant that patients were not seeking care for a period of time in Malawi [31]. In India, 17.3% of patients defaulted on their TB treatment and others consulted general practitioners and private pharmacies for treatment due to the difficulty in accessing healthcare facilities [38]. A survey by the Stop TB partnership found that in several countries, fear of contracting COVID-19 kept patients away from visiting clinics [47]. Likewise, in Malawi, fear and ignorance of COVID-19 meant that many healthcare personnel refused to see or treat anyone displaying symptoms resembling COVID-19 [31]. Moreover, staff were increasingly reluctant to handle any sputum samples or observe sputum collection. Furthermore, this was the case in Nigeria [32,38]. A lack of personal protective equipment (PPE) discouraged staff from attending to patients in many countries [31,47]. A survey by the Stop TB partnership found that staff at TB clinics observed a need for patients to be given nutritional support, as well as have their transportation costs covered for visiting healthcare facilities [47].

## 3.4.3. Recommendations and Adaptations of TB Services

Five studies from multiple authors including the WHO have detailed recommendations on how TB services can be improved during a pandemic in high burden settings. All of the studies agreed that the use of telemedicine can be leveraged for TB care. Medical triage and counselling should be conducted by telephone. Where possible, sputum collection should be conducted in a well-ventilated area at home and staff must be adequately protected when collecting the samples from patients [46]. The switch to oral and shorter treatment regimens [42], as well as the video-supported treatment would reduce the number of patients visiting health facilities [46]. Integrating TB and COVD-19 care,

such as testing and active case finding, could benefit the management of both diseases [42]. HIV care must also be integrated for countries with a high disease burden [42]. Patients with drug-susceptible TB should be provided with enough TB medication for the intensive phase and only return to the healthcare facility for an assessment. In addition, they need to switch to the continuation phase where sufficient medication is provided [43,45]. Patients with drug-resistant TB (DR-TB) must be switched to an oral treatment that lasts until the next scheduled visit, any patients exhibiting concerning iron levels or myelosuppression must be recalled by telephone [44,46]. Moreover, decentralizing the treatment collection has been encouraged [42,46]. Furthermore, there was an emphasis on strengthening primary care in order to help in managing the pandemic, by providing PHC workers with best practice training for COVID-19 [43,46]. This ensures that PHC facilities are equipped with enough staff who have access to PPE and provision of all chronic medication should be available for extended periods to reduce visits to health facilities [43]. Finally, all of the PHCs offering TB testing must follow the recommended infection prevention and control (IPC) measures, from the collection of samples until testing is conducted and the sample is disposed of in the laboratory [46]. It is not clear how many high burden countries have implemented these changes for their TB programs and how successful implementation has been. The following section explores examples of instances where TB services have been adapted.

## 3.4.4. Implementing the Recommendations for TB Services

Five of the included studies documented the changes made to the TB services in response to the COVID-19 pandemic. Outpatient departments in India and Portugal screened patients for COVID-19 before they were attended to. Crowd control was also maintained to ensure that social distance and IPC measures are upheld at all times [28,40]. The same center in Portugal did contact tracing by telephone and patients were only asked to come to the clinic if they had a positive screening after the phone call [28]. The oral treatment is now favored over injectables and treatment is administered in line with scheduled healthcare visits in India and the USA [32,40]. Those requiring intravenous treatments are administered by community nurses at home [40]. In several countries, treatment initiation is conducted in clinics, but all of the follow-ups are conducted by telephone, including any consultation with doctors, unless presenting with severe symptoms or treatment side effects [28,31,34,40]. TB clinics in New York, USA, have also begun giving patients daily reminders over the phone to ensure that they adhere to the treatments [28]. In cases where patients cannot utilize telehealth due to limitations in technology, then home visits are conducted on a case by case basis [32]. In Pakistan, general practitioners (GPs) who referred patients to TB centers were used to locate patients that could not be contacted during the pandemic. Moreover, their offices were used as a location where patients could fetch their medication [33]. Furthermore, certain provinces in Pakistan have mandated that private healthcare providers notify TB cases to national TB programs [34]. Healthcare workers are provided with the necessary PPE according to the risk of exposure and they work in shifts to avoid overcrowding [40]. In Pakistan, healthcare providers have been retrained in IPC and the correct use of PPE wherever necessary [34]. All of these adaptations are new and will need to be closely monitored throughout the pandemic to assess their sustainability and effectiveness. Furthermore, more data are needed on other high burden countries to see whether they have adapted TB services since the start of the pandemic.

## **4. Discussion**

This scoping review mapped the existing evidence of TB services at the PHC level in the COVID-19 era. The evidence was from a wide range of documentary sources, and most came from high TB burden regions of Pakistan, India, Nigeria, and South Africa. The bulk of the literature found was from the start of the pandemic. The findings show evidence that the COVID-19 pandemic had a negative effect on TB services, how patients and healthcare providers were impacted, as well as recommendations for adapting these

services and instances where recommendations had been implemented. Overall, the COVID-19 pandemic has negatively impacted TB services, users, and healthcare providers alike. The findings suggest that TB services were disrupted. In addition, the fear and stigma experienced by healthcare providers and patients likely led to a drop in TB case detection and the notifications seen during the first months of the pandemic. More evidence is needed on the steps taken to identify potentially undiagnosed and missed TB cases and how provider attitudes and patient experiences have improved, especially in high TB burden countries. Although the review has highlighted recommendations for enhancing TB services in high burden settings during the pandemic, only one TB endemic country had implemented these changes. Before COVID-19, countries were making strides towards achieving the SDG targets for TB; a record number of people had been treated including those with DR-TB; the annual number of missed TB cases had fallen below 3 million; and the TB preventative treatment had been prioritized in high burden settings [7]. However, this has likely changed since the start of the COVID-19 pandemic.

The findings of this scoping review show that the arrival of COVID-19 and the measures used to curb the spread drastically reduced the number of TB cases detected and notified, in sharp contrast to the numbers from the same period in previous years [29,32,36]. It further demonstrated how TB services were significantly disrupted and sidelined in response to a new public health emergency [9,32,37,45]. Moreover, the COVID-19 pandemic deterred health-seeking behaviors, hindered some patients from acquiring the TB treatment, and increased reluctance among healthcare workers to treat patients [31,37]. These results have created scenarios for TB cases to go undiagnosed. Furthermore, the fear of attending health facilities and the disruptions leading to their closure have likely interrupted TB treatment regimens, which could lead to treatment failure exacerbating disease transmission and development of drug resistance. These would be grim outcomes for global TB control efforts. The responses, uncovered by the review, mirror those experienced during the Ebola virus outbreak in West Africa, which increased preventable TB deaths over time [48]. Following the Ebola outbreak, TB, HIV, and malaria deaths exceeded those directly caused by the Ebola virus itself [49,50]. Similarly, the outbreak of the Middle East respiratory syndrome coronavirus (MERS-CoV) in Saudi Arabia harmed TB control efforts [51]. As a result of the Ebola outbreak [52], certain West African countries have implemented measures for epidemic response. However, it is unclear whether these have had any bearing on TB control during the COVID-19 era. The increase in TB mortality in the coming years due to the disruptions to health services has been foreshadowed by several studies [13,14,53]. These will be most evident in high burden settings unless swift action is taken to minimize the impact on health services, while simultaneously identifying, diagnosing, and treating any cases that are not from the start of the pandemic.

A study summarizing the effect of Ebola on TB services emphasized the importance of moving away from disease-specific national programmes to the holistic strengthening of health systems [48]. It further highlights how this kind of approach would not only assist with the management of infectious outbreaks, but ensure that disease control for other conditions is not compromised [48]. Two previous studies showed that China's and Saudi Arabia's prior coronavirus experience facilitated a better COVID-19 response than many other countries [54,55]. In contrast, this review did not find evidence of TB endemic countries adopting the lessons from previous epidemics. However, our findings present current recommendations for conducting TB services during the COVID-19 pandemic [42–44,46]. Although these are helpful, it would have been more beneficial if governments had adopted insights from past viral outbreaks. The review also demonstrates how the recent adaptations to TB services have been adopted in various countries [28,32–34,40]. However, only two of these were high TB burden countries [34,40]. Considering that many of these suggestions rely on the use of technology, their practicality for resource-limited settings remains to be seen. Therefore, high burden countries must continue to monitor the impact of COVID-19 on TB services and address these with evidence-based interventions.

## *4.1. Implication for Research*

Many of the included studies documented the situation at the start of the COVID-19 pandemic. Consequently, geographic areas with a lower incidence at the start of the pandemic were not a focus of this study. Thus, an assessment of TB services in these regions is needed for better insight into the global effect of the COVID-19 pandemic. Research on how TB services have fared throughout the pandemic, including peaks in COVID-19 cases and subsequent vaccination strategies, are also needed. The same can be said for provider and health seeker attitudes. This will facilitate the measurement of the impact on TB and allow appropriate mitigation action. Undiagnosed TB cases and interruption to treatment were other issues likely to have been caused by the pandemic. Health systems must be ready to receive and appropriately treat and retain these cases until treatment completion. Therefore, assessing the quality of TB services in high burden settings and providing context-specific adaptations based on the findings could benefit TB control programmes. Moreover, only four empirical studies were found and even these scored low in terms of methodological quality. Therefore, robust primary studies are required to inform evidence-based decisions and recommendations for TB services during pandemics. These studies should focus on strengthening TB case findings, diagnostics, and treatment services for COVID-19 and future pandemics.

## *4.2. Strengths and Limitations*

The scoping review employed a comprehensive database search that was not limited by language, publication or study design. In addition, the database search included a grey literature and repeated search in the database that retrieved the highest number of articles to maximize the number of studies found. The methodological quality of all the included primary studies was assessed and it was found that they ranged from a low to average quality. For this reason, the scoping review may not be appropriate to inform clinical practice, but does demonstrate a need for additional primary studies to be conducted with more methodological rigor. Moreover, the review retrieved evidence from the start of the COVID-19 pandemic. Therefore, certain geographic regions with initially low incidence rates were not covered. Furthermore, given the evolving nature of the pandemic, it is likely that latter phases including the emerging variants and vaccination control strategies have also impacted the TB service delivery.

## **5. Conclusions**

In this review, the TB services at the PHC level were disrupted by the COVID-19 pandemic. The potential for undiagnosed TB cases and treatment failure are among the biggest concerns caused by the pandemic. For the TB elimination goals to be met, PHC must be strengthened and ready with effective solutions to address the issues caused by the COVID-19 pandemic and use these for pandemic preparedness in the future.

**Author Contributions:** Conceptualization, T.D. and T.M.-T.; developing and conducting the search strategy, K.K.; screening, K.M., A.B., and Z.N.J.; writing—original draft, T.D.; writing—reviewing and editing, T.M.-T. and A.M.; supervision, A.M. and T.M.-T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This presentation/publication has been made possible by funding from UNICEF and with support from Future Africa, the University of Pretoria.

**Data Availability Statement:** The data for the scoping review was obtained through secondary data analysis, all data supporting the conclusions of this scoping review are available through the reference list.

**Acknowledgments:** Cheryl Tosh for editing.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.




**Table A1.** *Cont.*


**Table A1.** *Cont.*

## **References**


## *Editorial* **Diagnostics Literacy Advocacy Model for Vulnerable Populations**

**Tivani P. Mashamba-Thompson**

Faculty of Health Sciences, University of Pretoria, Pretoria 0002, South Africa; tivani.mashamba-thompson@up.ac.za

**Abstract:** Evidence shows that vulnerable populations have lower levels of health literacy, resulting in poor health-seeking behavior and poor uptake of diagnostics. Being health literate promotes health care-seeking behavior and improves engagement with diagnostic services. In this editorial, I define health literacy in the context of access to technology for enabling disease screening, diagnosis and linkage to care. I refer to health literacy in this context as diagnostics literacy. The COVID-19 pandemic has taught us that vulnerable populations are disproportionately disadvantaged by the disruptive measures put in place to control the spread of the virus. Many vulnerable populations are still experiencing short-and longer-term socio-economic consequences. I propose a multi-level diagnostics literacy advocacy model to help improve diagnostic uptake among vulnerable populations.

**Keywords:** vulnerable populations; literacy; diagnostics; advocacy

In response to the COVID-19 pandemic, researchers have scaled up the development of new diagnostics [1]. Effective diagnostics improve detection of SARS-CoV-2 infected patients and the overall surveillance of the COVID-19 pandemic. Recent literature shows that 47% of the world's population has poor access to diagnostics and are failing to achieve the United Nations General Assembly (UNGA) COVID-19 testing targets of one test per 1000 people per day [2]. Improving access to diagnostics by removing barriers may reduce annual premature deaths by 1.1 million (2.5% of total annual deaths) and morbidity by 38.5 million (1.8% of all conditions) annual disability-adjusted life-years lost in low-income and middle-income countries (LMICs) [3]. Resource-constrained settings with limited laboratory infrastructure also have poor access to diagnostics. Additionally, the SARS-CoV-2 virus spreads inequitably through vulnerable populations with limited access to healthcare, resulting in higher rates of infection and complications. Vulnerable population groups have also been disproportionately disadvantaged by disruptive measures put in place to stop the spread of the virus. Many people who relied on small day-to-day earnings lost income required to meet basic needs such as housing and food. Women and girls who were caring for family members lost educational and professional opportunities and will continue to experience long-term socio-economic consequences. Vulnerable population groups may have lower levels of health literacy, resulting in sub-optimal health-seeking behavior and poor diagnostics uptake.

Improving access to diagnostics is a global health priority and we need to explore factors that limit access and use of available diagnostics in settings that have limited access to laboratory infrastructure. Health literacy is known to influence the use of health services [4]. Health literacy is a major determinant of health outcomes and is imperative to global health [5]. Health literacy is defined as a set of skills that allows patients to control their own well-being, allows them to make smart healthcare choices, improves patients' communication with healthcare workers and gives them the information to advocate for themselves in healthcare settings [6]. High levels of health literacy improve access to health care services, including diagnostics [7]. Here, I define health literacy in the context of accessing technology to enable disease screening, diagnosis and linkage to care. I refer to

**Citation:** Mashamba-Thompson, T.P. Diagnostics Literacy Advocacy Model for Vulnerable Populations. *Diagnostics* **2022**, *12*, 716. https:// doi.org/10.3390/diagnostics12030716

Received: 10 March 2022 Accepted: 14 March 2022 Published: 15 March 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

health literacy in this context as diagnostics literacy (DL). Diagnostic literacy encompasses a broad range of factors that are closely related to health promotion, including culture, individual empowerment, community development, media and numeracy. We urgently need to implement a multi-level DL advocacy model (Figure 1) to improve diagnostics uptake among underserved populations.

## **Demographics**

**Figure 1.** Multi-level diagnostics literacy advocacy model.

## **The Proposed Multi-Level Diagnostics Literacy Advocacy Model**


Diagnostic literacy should be improved in vulnerable populations. Well-structured advocacy strategies should be prioritized. Improved DL is likely to lead to increased up take of diagnostics services. Messages can be disseminated using channels tailored to communities to enhance and promote disease diagnosis-seeking behavior. Communities should be fully engaged to ensure that the messages become engrained and translate to health-seeking behavior.

**Funding:** This research received no external funding.

**Acknowledgments:** The author would like to extend her appreciation to Cheryl Tosh for editorial services.

**Conflicts of Interest:** The author declares no conflict of interest.

## **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Diagnostics* Editorial Office E-mail: diagnostics@mdpi.com www.mdpi.com/journal/diagnostics

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-4866-1