Key messages
Using digital tools to help healthcare workers track patient information and make decisions has mixed results. These approaches may lead to benefits in some situations but may make little or no difference in others. We often do not know the full impact because there are not enough studies, or the studies do not provide clear answers.
What did we want to find out?
This review examined whether digital tools can help healthcare workers collect better patient information and provide better care compared to traditional methods, like paper records or no tools at all. The focus was on primary healthcare settings, such as clinics and community health programmes.
What are digital tracking systems?
Digital tracking systems help healthcare workers collect and store patient information over time. This can help them provide better care and allow governments to design better health services.
Patient information is often stored in electronic medical records in wealthier countries. However, healthcare workers in many low- and middle-income countries still use paper records or have limited access to digital tools. Mobile devices like tablets and smartphones can make it easier for healthcare workers in these settings to use digital systems to track patient information more accurately and securely.
What other tools are included?
Decision support tools: Digital tracking systems are sometimes combined with decision support tools. These tools guide healthcare workers in analysing patient data and deciding on the best treatment.
Targeted client communication: Digital tracking systems can also be combined with systems for communicating with patients and clients. These tools allow healthcare workers to send patient messages, such as appointment reminders, medication schedules, or health advice.
What did we find?
The review included 18 studies. Seventeen were from low- and middle-income settings where electronic medical record systems are rare or just being rolled out. Most studies focused on healthcare workers using mobile devices, particularly for maternal and child health or managing chronic diseases.
When primary healthcare workers use digital tracking systems combined with decision support tools, targeted client communication, or both, compared to usual care, the results are mixed.
Positive effects: These tools may have some benefits. For example, they may help healthcare workers follow guidelines better, keep more accurate records and provide better quality of care. They may also improve patient outcomes. For example, some mothers may have fewer low-birthweight babies or may start breastfeeding earlier. Some patients may also have improved quality of life or be more consistent with taking medications for chronic conditions such as stroke. However, these positive effects may be small.
Little to no impact: In other cases, the tools may make little or no difference to healthcare worker behaviour or patient health.
Uncertainty: In many cases, the results are unclear because there are not enough studies or the studies do not provide strong evidence.
What are the limitations of these findings?
We have little confidence in the findings because there were not many studies; some studies were small, and others had design problems that made their findings less certain.
How up‐to‐date is this review?
The review examined studies conducted between 2000 and November 2022.
Digital tracking may improve primary care workers' ability to follow recommended antenatal and chronic disease practices, quality of patient records, patient health outcomes and service use. However, these interventions led to small or no outcome differences in most studies.
Digital tracking on mobile devices, combined with clinical decision support systems and targeted client communication, can facilitate service delivery and potentially improve outcomes.
To assess the effects of using a mobile device to track service use when combined with clinical decision support (Tracking + CDSS), with targeted client communications (Tracking + TCC), or both (Tracking + CDSS + TCC).
Cochrane CENTRAL, MEDLINE, Embase, Ovid Population Information Online (POPLINE), K4Health and WHO Global Health Library (2000 to November 2022).
Randomised and non-randomised trials in community/primary care settings.
Participants: primary care providers and clients
Interventions:
1. Tracking + CDSS
2. Tracking + TCC
3. Tracking + CDSS + TCC
Comparators: usual care (without digital tracking)
Two authors independently screened trials, extracted data and assessed risk of bias using the RoB 1 tool. We used a random-effects model to meta-analyse data producing risk differences (RD), risk ratios (RR), or odds ratios (OR) for dichotomous outcomes and mean differences (MD) for continuous outcomes. Evidence certainty was assessed using GRADE.
We identified 18 eligible studies (11 randomised, seven non-randomised) conducted in Bangladesh, China, Ethiopia, India, Kenya, Palestine, Uganda, and the USA. All non-randomised studies had a high risk of bias.
These results are from randomised studies.
'Probably/may/uncertain' indicates 'moderate/low/very low' certainty evidence.
Tracking + CDSS
Relating to antenatal/ postnatal care:
Providers' adherence to recommendations
May slightly increase home visits in the week following delivery (2 studies, 4531 participants; RD 0.10 [0.07, 0.14])
May slightly increase counselling for initiating complementary feeding (2 studies, 4397 participants; RD 0.12 [0.08, 0.15])
May slightly increase the mean number of home visits in the month following delivery (1 study, 3023 participants; MD 0.75 [0.47, 1.03])
Uncertain effect on home visits within 24 hours of delivery
Clients' health behaviours
May slightly increase skin-to-skin care (1 study, 1544 participants; RD 0.05 [0.00, 0.10])
May slightly increase early breastfeeding (2 studies, 4540 participants; RD 0.08 [0.05, 0.12])
Uncertain effects on applying nothing to the umbilical cord, taking ≥ 90 iron-folate tablets during pregnancy, exclusively breastfeeding for six months, delaying the newborn's bath at least two days and Kangaroo Mother Care.
Clients' health status
May reduce low birthweight babies (1 study, 3023 participants; RR 0.53 [0.38, 0.73])
May increase infants with pneumonia or fever seeking care (1 study, 3470 participants; RR 1.13 [1.03, 1.24])
Uncertain effects on stillbirths, neonatal and infant deaths, or testing positive for HIV during antenatal testing
Tracking + TCC
Clients' health status
In stroke patients over 12 months:
May slightly increase blood pressure (BP) medication adherence (1 study, 1226 participants; RR 1.10 [1.00, 1.21])
May reduce deaths (1 study, 1226 participants; RR 0.52 [0.28, 0.96])
May slightly reduce systolic BP (1 study, 1226 participants; MD -2.80 mmHg [-4.90, -0.70])
May slightly improve EQ-5D scores (1 study, 1226 participants; MD 0.04 [0.02, 0.06])
May reduce stroke hospitalisations (1 study, 1226 participants; RR 0.45 [0.32, 0.64]).
Tracking + CDSS + TCC
Providers' adherence to recommendations
Probably increases guideline adherence for antenatal screening and management of anaemia (1 study, 10,502 participants; OR 1.88 [1.52, 2.32]), diabetes (1 study, 8669 participants; OR 1.45 [1.14, 1.84}), hypertension (1 study, 15,555 participants; OR 1.62 [1.29, 2.04]) and probably leads to lower adherence for abnormal foetal growth (1 study, 1165 participants; OR 0.59 [0.37, 0.95]).
May slightly increase nevirapine prophylaxis in infants of HIV+ve mothers (1 study, 609 participants; OR 1.75 [0.73, 4.19])
Data quality
In pregnant women (1 study, 6367 participants), tracking + CDSS + TCC:
Probably slightly reduces missing data for haemoglobin (RR 0.77 [0.71, 0.84]) but slightly more for BP at delivery (RR 1.16 [1.08, 1.24])
May have little or no effect on missing data on gestational age (RR 0.96 [0.81, 1.14]) or birthweight (RR 0.90 [0.77, 1.04])
Clients' health behaviour
May have little or no effect on being on anti-retroviral therapy at delivery (1 study, 438 participants; OR 1.41 [0.81, 2.45]) or exclusive breastfeeding for six months (1 study, 695 participants; OR 1.74 [0.95, 3.17]) in HIV+ve mothers
Uncertain effects on physical activity in high cardiovascular-risk adults
Clients' health status
May reduce the number of deaths in patients with hypertension and diabetes (1 study, 3698 participants; OR 0.61 [0.35, 1.06])
May reduce new cardiovascular events in high-cardiovascular risk adults over 6-18 months (1 study, 8642 participants; OR 0.58 [0.42, 0.80})
May slightly decrease in antenatal women severe hypertension, but the confidence interval includes both a decrease and increase (1 study, 6367 participants; OR 0.61 [0.27, 1.37])
In women receiving antenatal care (1 study, 6367 participants), tracking + CDSS + TCC maymake little or no difference to adverse pregnancy outcomes (OR 0.99 [0.87, 1.12]), moderate or severe anaemia (OR 0.82 [0.51, 1.31]), or large-for-gestational-age babies (OR 1.06 [0.90, 1.25]).
In adults with hypertension or diabetes (1 study, 3324 participants), tracking + CDSS + TCC maymake little or no difference to HbA1c (MD 0.08 [-0.27, 0.43]), total cholesterol (MD -2.50 [-7.10, 2.10]), 10-year cardiovascular risk (MD -0.40 [-2.30, 1.50]), tobacco use (MD-0.05 [-0.47, 0.37]), alcohol use (MD 0.70 [-3.70, 5.10]), or PHQ-9 (MD -1.60 [-4.40, 1.20]).
Uncertain effects on maternal or infant mortality before the baby reaches 18 months in HIV-positive mothers, patients who achieve optimal BP, BP controlled at five years, diastolic or systolic BP, body mass index, fasting glucose and quality of life in adults with hypertension or diabetes
Client service utilisation
May have little or no effect on missed early infant diagnosis visits (1 study, 1183 participants; OR 0.92 [0.63, 1.35]).
Uncertain effects on linkage to care
Client satisfaction
Probably increases slightly the number of adults with hypertension or diabetes reporting "slightly/much better" change in the quality of care (1 study, 3324 participants; RR 1.02 [1.00, 1.03]).
No studies evaluated time between presentation and appropriate management, timeliness of receiving/accessing care, provider acceptability/satisfaction, resource use, or unintended consequences.