Effect of decision-support tools on mobile phones on primary health care

What was the aim of this review?

In this Cochrane Review, we aimed to find out if primary (community) healthcare workers using decision-support tools on mobile phones or other mobile devices give better quality care. We looked for studies where researchers compared a decision-support tool used on mobile phones to routine practice where there may be no guidance or some guidance in a paper format. We searched for studies conducted from 1 January 2000 to 9 October 2020. We found eight studies.

Key messages

We do not know if decision-support tools used on mobile devices make primary healthcare workers better at following recommended practice. The evidence is not clear about the effects of these tools on patients' and clients' behaviour and on their health. We need more and better research to assess these issues.

What was studied in the review?

In many settings, patients receive low-quality care. This is often because they live in poor or rural settings with few healthcare workers, or because healthcare workers do not have enough supplies, equipment, or proper training. Healthcare workers may struggle to stay up-to-date or may not have enough time to make the right decisions, which can result in poor quality of care for patients.

Decision-support tools may help address some of these problems. A decision-support tool helps the healthcare worker think through what he or she knows about the patient. The tool then helps guide the healthcare worker to the right decision for that patient. Designing decision-support tools that can be used on mobile phones or other mobile devices such as tablets and personal digital assistants (PDAs) can make these tools easier to use and keep up-to-date.

The main aim of our review was to find out if healthcare workers using decision-support tools on mobile phones give better healthcare than healthcare workers using decision-support tools that are not on mobile phones or that use no decision-support tools. We looked at the use of these tools in primary healthcare settings only.

What were the main results of the review?

We found eight relevant studies. Three studies were carried out in the USA and five studies in India, China, Guatemala, Ghana, and Kenya. These studies showed that when primary healthcare workers use decision-support tools on mobile phones:

– we do not know if they are better at following recommended clinical practice, because the quality of this evidence was very low;

– there was no clear pattern of a positive or negative effect on patients' or clients' behaviour and on their health;

– this may slightly improve patients’ satisfaction with medical information;

– we do not know if this approach led primary healthcare workers to manage people’s health issues more quickly because we found no studies that measured this. We also found no studies that explored the effect on healthcare worker satisfaction, resource use, or whether this approach had any unintended consequences (e.g. harms).

How up-to-date is this review?

We searched for studies published up to October 2020.

Authors' conclusions: 

We are uncertain about the effectiveness of mobile phone-based decision-support tools on several outcomes, including adherence to recommended practice. None of the studies had a quality of care framework and focused only on specific health areas.   We need well-designed research that takes a systems lens to assess these issues.

Read the full abstract...

The ubiquity of mobile devices has made it possible for clinical decision-support systems (CDSS) to become available to healthcare providers on handheld devices at the point-of-care, including in low- and middle-income countries. The use of CDSS by providers can potentially improve adherence to treatment protocols and patient outcomes. However, the evidence on the effect of the use of CDSS on mobile devices needs to be synthesized. This review was carried out to support a World Health Organization (WHO) guideline that aimed to inform investments on the use of decision-support tools on digital devices to strengthen primary healthcare.


To assess the effects of digital clinical decision-support systems (CDSS) accessible via mobile devices by primary healthcare providers in the context of primary care settings.

Search strategy: 

We searched CENTRAL, MEDLINE, Embase, Global Index Medicus, POPLINE, and two trial registries from 1 January 2000 to 9 October 2020. We conducted a grey literature search using mHealthevidence.org and issued a call for papers through popular digital health communities of practice. Finally, we conducted citation searches of included studies.

Selection criteria: 

Study design: we included randomized trials, including full-text studies, conference abstracts, and unpublished data irrespective of publication status or language of publication. 

Types of participants: we included studies of all cadres of healthcare providers, including lay health workers and other individuals (administrative, managerial, and supervisory staff) involved in the delivery of primary healthcare services using clinical decision-support tools; and studies of clients or patients receiving care from primary healthcare providers using digital decision-support tools.

Types of interventions: we included studies comparing digital CDSS accessible via mobile devices with non-digital CDSS or no intervention, in the context of primary care. CDSS could include clinical protocols, checklists, and other job-aids which supported risk prioritization of patients. Mobile devices included mobile phones of any type (but not analogue landline telephones), as well as tablets, personal digital assistants, and smartphones. We excluded studies where digital CDSS were used on laptops or integrated with electronic medical records or other types of longitudinal tracking of clients.

Data collection and analysis: 

A machine learning classifier that gave each record a probability score of being a randomized trial screened all search results. Two review authors screened titles and abstracts of studies with more than 10% probability of being a randomized trial, and one review author screened those with less than 10% probability of being a randomized trial. We followed standard methodological procedures expected by Cochrane and the Effective Practice and Organisation of Care group. We used the GRADE approach to assess the certainty of the evidence for the most important outcomes.

Main results: 

Eight randomized trials across varying healthcare contexts in the USA,. India, China, Guatemala, Ghana, and Kenya, met our inclusion criteria. A range of healthcare providers (facility and community-based, formally trained, and lay workers) used digital CDSS. Care was provided for the management of specific conditions such as cardiovascular disease, gastrointestinal risk assessment, and maternal and child health. The certainty of evidence ranged from very low to moderate, and we often downgraded evidence for risk of bias and imprecision.

We are uncertain of the effect of this intervention on providers' adherence to recommended practice due to the very low certainty evidence (2 studies, 185 participants). The effect of the intervention on patients' and clients' health behaviours such as smoking and treatment adherence is mixed, with substantial variation across outcomes for similar types of behaviour (2 studies, 2262 participants). The intervention probably makes little or no difference to smoking rates among people at risk of cardiovascular disease but probably increases other types of desired behaviour among patients, such as adherence to treatment. The effect of the intervention on patients'/clients' health status and well-being  is also mixed (5 studies, 69,767 participants). It probably makes little or no difference to some types of health outcomes, but we are uncertain about other health outcomes, including maternal and neonatal deaths, due to very low-certainty evidence. The intervention may slightly improve patient or client acceptability and satisfaction (1 study, 187 participants). We found no studies that reported the time between the presentation of an illness and appropriate management, provider acceptability or satisfaction, resource use, or unintended consequences.