A literature review was conducted to identify the barriers to assessing body temperature in low-resource settings. In order to clarify and quantify the relationship between fever diagnostics and a health care system, Uganda was selected as a representative country of the sub-Saharan African region. Publications were retrieved from Google Scholar and PubMed using the following keywords: fever and Uganda, barrier and febrile treatment, thermometer and diagnosis, drug shop, rural Uganda, healthcare, measuring body temperature, and misdiagnosis. An additional keyword, perception, was used after retrieving the publications from the first search. Simultaneously, related articles were searched based on data extracted from citation indices. Articles were selected if they included qualitative and/or quantitative studies that identified barriers to assess the body temperature in resource-constrained environments, especially in Uganda. Articles were excluded if they only focused on specific countries in low-to-middle-income economies excluding Uganda.

The five As of access to care by Penchansky and Thomas [48,49] were used to categorize the barriers identified in the searched literature. Characteristics and expectations of both health care providers and their clients were grouped into the five As: accessibility, availability, acceptability, affordability, and accommodation. This framework was selected among others for this study given its extensive use in the field of health care [48,50-53], its degree of detail, and its comprehensiveness regarding the different health care service users. Although the aim of this study was not to compare existing frameworks, the following description highlights the most relevant considerations for the authors’ choice.

In the five As framework, accessibility refers to the geographic distribution of health care facilities. Availability relates the existing quantity of resources (ie, personnel and technology) with the ones required to meet the demands of the people. In the framework of Peters et al [50], for example, these two dimensions are merged and are therefore less adequate in circumstances where incomplete or unsuitable health care facilities are located nearby health care seekers. Affordability relates the direct and indirect charges related to health care services to the ability and willingness of health care seekers to pay them. Acceptability refers to the inherent characteristics of the system in place regarding genre, ethnicity, and social class, for example, and is often susceptible to mutual social and cultural appraisals. Finally, accommodation is determined by the extent to which the offered services are adjusted to match the client’s access capacity (eg, hours of operation and people's ability to receive treatment without prior appointments) [48]. This aspect, in particular, points toward an interesting service design component and the consideration that systems can purposefully be designed to adjust to the lifestyle of health care seekers that other frameworks do not include.

The two latter aspects are often merged into one dimension. In Prahalad’s innovations in the bottom of the pyramid [53], these two aspects combined are renamed awareness of providers. In Peters et al [50], acceptability is only described from health provider’s perspective. The 4 As framework of the World Health Organization [5] is focused on medical equipment. Accommodation and acceptability are described as technical appropriateness to context. In McIntyre et al [51] and in Grimes [52], access barriers to health care in low- and middle-income countries are categorized into only three dimensions: acceptability, affordability, and availability [51]. Accessibility and availability are merged and defined as “being at the right place, at the right time.” Accommodation and acceptability are seen as a corresponding dimension between a participant’s expectations and the services provided.

A total of 37 articles were included and reviewed. These include 25 studies that relate to treatment of febrile illnesses, of which seven address fever diagnostics and three address health care services in Uganda. Also included in the literature were four studies that looked at medical devices in low-resource settings and two studies that addressed more generally the barriers to accessing health care in low-resource settings. We identified 11 main barriers to accessing and receiving adequate fever diagnostics that were divided into the five categories (see Table 1) [3,4,31,33,40,44,54-63]. They will be discussed in detail in the following sections.

The difficulty and delay in accessing treatment of febrile illnesses is attributed to a large extent to the physical distance between health care providers and health care seekers. The physical distance to health care providers influences people’s choices of health care providers when seeking care for febrile illnesses. This mostly affects people living in rural areas in Uganda, where the majority of the population (84.4%) lives [54]. According to the definition from the Ugandan government, the health care sector can be divided into the public sector and private sector [56] (see Textboxes 1 and 2).

The private health sector is categorized into private for-profit, private nonprofit, and informal providers. Drug shops categorized into private for-profit account for the largest proportion of all facilities in the private health sector in all districts except Kampala, where more clinics than drug shops can be found [56]. Public facilities, which include hospitals and health centers (II, III, and IV), make up 54.8% of the total Ugandan health care facilities, while 28.5% are private for-profit and 16.7% are private nonprofit [64] (see Figure 1). However, the distribution of health care facilities in rural areas is significantly different from the distribution of health care facilities in Uganda as a whole. While more than half of all health care providers in Uganda are from the public sector, the public sector accounts for only 18.6% in rural districts where the majority of health care providers are private for-profit (74.5%) (see Figure 1).

Regarding the distribution of care sought by people with febrile symptoms, 31.1% of people sought care from a health care provider. Among health providers, excluding traditional healers, the main providers visited by people with febrile symptoms were private for-profit providers (51.8%), followed by public sector (39.8%) and private nonprofit providers (8%) (see Figure 2). Despite other treatment options being available in the community, the majority of people suffering from febrile symptoms treated their febrile illnesses by themselves at home (43.5%) or took no action (22.4%) [56]. The main reason given for visiting private providers instead of public health care providers is the convenience of location (ie, proximity) (see Figure 3). This may be explained by the fact that the great majority of health care facilities are private for-profit (74.5%) in rural Ugandan districts (see Figure 1). The distance to health care facilities also impacts the timing of care. Delay of treatment for fever occurred less among the people who perceived the distance between their home and the health care provider to be less than 1 km compared to those who perceived it to be more than 1 km [44].

Among the health care facilities, public facilities are perceived as having qualified and experienced health care providers by people seeking care for fever [55]. However, the government health sector is underresourced and understaffed and primary diagnostic equipment is frequently missing (see Figure 4). Compared to the private sector, where more than half of the private for-profit and nonprofit facilities are equipped with thermometers, public facilities were the worse equipped with thermometers among the formal health care facilities [56]. Furthermore, even though thermometers are available, a chronic understaffing problem in the public sector leads clinicians to routinely and inadequately assess patients’ body temperature by placing a hand on their foreheads, versus utilizing thermometers, when it is peak time in the waiting room [57].

Regarding staff qualification, the private sector is invariably inferior to the public sector [3]. However, even though health care providers in the public sector were perceived as experienced, only 3 in 10 public health care professionals were able to diagnose 4 out of 5 very common illnesses (ie, malaria with anemia, acute diarrhea with severe dehydration, pneumonia, pulmonary tuberculosis, and diabetes mellitus). Among the most common, malaria with anemia was the least likely to be diagnosed correctly and only 9% of the cases were recommended the appropriate treatment [3]. Fever is more likely to be assessed by tactile measurement (ie, placing the palm or back of the hand on the forehead) than with a fever thermometer and the changes of fever over time are observed by patients or parents of child patients [31]. The absence of a fever thermometer at home hampers and delays the treatment of fever. Clinicians and nurses claim that people should have thermometers ready at home to quickly and objectively assess body temperature and be able to deduce how serious an illness may be [57]. Due to medical resource constrains in the public sector, the private for-profit facilities perform a key role in the supply of medicine. The main providers (83%), where febrile patients purchase medicine, are the private for-profit facilities and the second-major provider (10%) is the informal sector (see Figure 5) [56]. Informal providers are numerous, nearby, and more consumer oriented [65]. However, the private sector’s knowledge and quality of treatment at drug shops are recognizably limited [4]. While personnel with good technical skills was the main reason given for choosing public providers (45%), 26.4% of people perceived that the private for-profit providers had personnel with good technical skills (see Figure 3). In addition, even though 85% of the public health care facilities were inspected by local authorities monthly or quarterly, only half of the private for-profit facilities (54%) were inspected monthly or quarterly and 36% were never inspected at all [56].

Besides the associated cultural beliefs regarding the subjectivity of fever itself, there are relevant acceptability aspects about how and when fever is measured among community social networks. Social networks and common practices in the communities play an important role in fever-related health care decisions. Nsungwa-Sabiiti et al [66] describe how mothers are often reluctant to seek professional health care for their ill children at an early stage [58]. Feeling unwell with any kind of fever symptom is perceived as the most important disease in their community and is believed to be caused by something you ate or drank, environmental conditions, mosquitoes, and being a symptom of other diseases. There was consensus among the members of the community that care for febrile symptoms is to be sought from the informal sector before visiting the formal sector. The health care facilities are visited as a definitive way to care for febrile people after treatment with herbs and medicine purchased from the shops [67].

Technologies and services do not accommodate the needs, expectations, or habits of health care seekers in several ways. First, the reading of the thermometer, as it is designed, is often not understood. This may be due to the multiplicity of different meanings people associate with fever or a febrile condition. In a setting where there is little or poor information available about the required follow-up of fever with regard to required dosages of medication and risks associated with diseases, the diagnostic information provided by the thermometer does not match the semantics associated with fever [57,66]. Second, this is not limited to households. The lack of knowledge in health care services to manage nonmalaria febrile illnesses results in health workers treating patients that have a negative malaria test result with antimalarial medicines [63]. As the patients expect to receive care for their symptoms, it is essential to provide appropriate management and information of febrile symptoms to those people who do not have a malaria infection. Finally, another limitation related to the diagnostic information provided by most fever thermometers is related to fever kinetics. The reporting of patients’ fever kinetics (ie, the progress of fever over time versus a punctual measurement) is essential for an adequate and accurate diagnosis. Since people often seek health care with delay, it is important for health care providers to know if the progress of fever based on memory is reliable [31]. Despite the accuracy of digital fever thermometers, since normal body temperature has individual variations, patients and parents need to know the patients’ baseline morning temperature to be able to judge an increased temperature as fever [7].

While the cost of treatment was a relatively minor determinant among a range of barriers to assess primary fever diagnosis compared to accessibility, the financial challenge is still one of the critical barriers and a concern for people seeking treatment for fever. The socioeconomic status of households has an effect on the timing of care seeking. Figure 6 shows the percentage of febrile children in Uganda that effectively sought timely care in different socioeconomic quintiles. Children with the lowest socioeconomic status were more likely to receive delayed care [44]. Although public health care services in Uganda are meant to be free, most patients have to pay for the treatment they receive and for the costs implied in transportation. Konde-Lule et al [56] demonstrated that in the public sector, half of the clients were charged for health services and paid an average of UGX 5381 (about €1.4). As expected, the majority of clients seeking care in private for-profit (86.5%) and nonprofit (84%) facilities were charged. The lowest average amount for health services is in private for-profit facilities where they cost UGX 4626 (about €1.2); the highest average amount for health services is in private nonprofit facilities where they cost UGX 7647 (about €2). The average monthly income of employees in 2013 was UGX 491,000 (about €128). Figure 7 illustrates the share of household expenditure by item group; health care expenditure accounts for 5% of total household expenditures [64].

This study presents an outline of the barriers of access to fever diagnostics in low-resource settings. This study also discusses an approach that may lead to an improved fever thermometer and help to reveal opportunities for innovative, complementary, and holistic initiatives to improve diagnostics of fever-related illnesses. On basis of the reviewed literature focused on sub-Saharan Africa, three complementary considerations were proposed that potentially have an impact in how fever diagnostics are designed and implemented in low-resource health care systems. Firstly, the fever diagnostics journey shows the involvement of people in the different phases of diagnostics, from awareness to monitoring and follow-up. Secondly, within the same health care system, there are different users of a fever thermometer for whom the conditions of access to fever diagnostics also differ. And thirdly, these different users have different needs regarding the information that is offered. The health care system in Uganda, as in other sub-Saharan countries, is greatly divided between public and private providers, and it is clear that the choices available for communities in low-resource settings are limited. In order to improve the overall access to fever diagnostics in these settings it is important to look into the specific and potential roles and needs that the different users may have. Needs related to fever diagnostics may include access to information about fever, information about its meaning and that of other illnesses, and clinical guidelines for handling and follow-up through appropriate channels. They may also include the need for appropriate thermometers and decision-making support. The involvement of health care professionals at all levels, community health workers, patients, and drug shop owners in a product-service system design approach may contribute to a more inclusive and holistic tackling of fever diagnostics.

The outcomes of this research are currently being used as direct input for the development of a new context-based product-service system for fever diagnostics in East Africa.