Limits...
Development and validation of a clinical and computerised decision support system for management of hypertension (DSS-HTN) at a primary health care (PHC) setting.

Anchala R, Di Angelantonio E, Prabhakaran D, Franco OH - PLoS ONE (2013)

Bottom Line: Software validation and piloting was done in field, wherein the virtual recommendations and advice given by the DSS were compared with two independent experts (government doctors from the non-participating PHC centers).Receiver operator curve (ROC) showed a good accuracy for the DSS, wherein, the area under curve (AUC) was 0.848 (95% CI: 0.741-0.948).Sensitivity and specificity of the DSS were 83.33 and 85.71% respectively when compared with independent experts.

View Article: PubMed Central - PubMed

Affiliation: Department of Public Health & Primary Care, University of Cambridge, Cambridge, United Kingdom ; Public Health Foundation of India, New Delhi, India.

ABSTRACT

Background: Hypertension remains the top global cause of disease burden. Decision support systems (DSS) could provide an adequate and cost-effective means to improve the management of hypertension at a primary health care (PHC) level in a developing country, nevertheless evidence on this regard is rather limited.

Methods: Development of DSS software was based on an algorithmic approach for (a) evaluation of a hypertensive patient, (b) risk stratification (c) drug management and (d) lifestyle interventions, based on Indian guidelines for hypertension II (2007). The beta testing of DSS software involved a feedback from the end users of the system on the contents of the user interface. Software validation and piloting was done in field, wherein the virtual recommendations and advice given by the DSS were compared with two independent experts (government doctors from the non-participating PHC centers).

Results: The overall percent agreement between the DSS and independent experts among 60 hypertensives on drug management was 85% (95% CI: 83.61-85.25). The kappa statistic for overall agreement for drug management was 0.659 (95% CI: 0.457-0.862) indicating a substantial degree of agreement beyond chance at an alpha fixed at 0.05 with 80% power. Receiver operator curve (ROC) showed a good accuracy for the DSS, wherein, the area under curve (AUC) was 0.848 (95% CI: 0.741-0.948). Sensitivity and specificity of the DSS were 83.33 and 85.71% respectively when compared with independent experts.

Conclusion: A point of care, pilot tested and validated DSS for management of hypertension has been developed in a resource constrained low and middle income setting and could contribute to improved management of hypertension at a primary health care level.

Show MeSH

Related in: MedlinePlus

Architectural diagram for the DSS system.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC3818237&req=5

pone-0079638-g002: Architectural diagram for the DSS system.

Mentions: Algorithms were developed by the software developers (Data Template, Bangalore, India) to help build the inferential engine base of the DSS. Medical language and software coding and machine language development were done by the medical software developers (Data Template) using "open source" platforms (JAVA and MySQL). Figure 2 details the architecture of the built DSS. Further details are mentioned in File S1. Caution was taken to ensure that the prepared ‘scenarios’, ‘risk stratifications’, ‘drug algorithms’ mirrored the Indian Hypertension II guidelines. The prepared “rules and logic” sheets [data collection form, drug indications, class of drugs and dosages, drug algorithms, undesirable combinations, risk stratification, referral scenarios and lab investigations (where possible)] were reviewed independently by two physician experts in the management of hypertension (government doctors from the non-participating PHC centers).


Development and validation of a clinical and computerised decision support system for management of hypertension (DSS-HTN) at a primary health care (PHC) setting.

Anchala R, Di Angelantonio E, Prabhakaran D, Franco OH - PLoS ONE (2013)

Architectural diagram for the DSS system.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC3818237&req=5

pone-0079638-g002: Architectural diagram for the DSS system.
Mentions: Algorithms were developed by the software developers (Data Template, Bangalore, India) to help build the inferential engine base of the DSS. Medical language and software coding and machine language development were done by the medical software developers (Data Template) using "open source" platforms (JAVA and MySQL). Figure 2 details the architecture of the built DSS. Further details are mentioned in File S1. Caution was taken to ensure that the prepared ‘scenarios’, ‘risk stratifications’, ‘drug algorithms’ mirrored the Indian Hypertension II guidelines. The prepared “rules and logic” sheets [data collection form, drug indications, class of drugs and dosages, drug algorithms, undesirable combinations, risk stratification, referral scenarios and lab investigations (where possible)] were reviewed independently by two physician experts in the management of hypertension (government doctors from the non-participating PHC centers).

Bottom Line: Software validation and piloting was done in field, wherein the virtual recommendations and advice given by the DSS were compared with two independent experts (government doctors from the non-participating PHC centers).Receiver operator curve (ROC) showed a good accuracy for the DSS, wherein, the area under curve (AUC) was 0.848 (95% CI: 0.741-0.948).Sensitivity and specificity of the DSS were 83.33 and 85.71% respectively when compared with independent experts.

View Article: PubMed Central - PubMed

Affiliation: Department of Public Health & Primary Care, University of Cambridge, Cambridge, United Kingdom ; Public Health Foundation of India, New Delhi, India.

ABSTRACT

Background: Hypertension remains the top global cause of disease burden. Decision support systems (DSS) could provide an adequate and cost-effective means to improve the management of hypertension at a primary health care (PHC) level in a developing country, nevertheless evidence on this regard is rather limited.

Methods: Development of DSS software was based on an algorithmic approach for (a) evaluation of a hypertensive patient, (b) risk stratification (c) drug management and (d) lifestyle interventions, based on Indian guidelines for hypertension II (2007). The beta testing of DSS software involved a feedback from the end users of the system on the contents of the user interface. Software validation and piloting was done in field, wherein the virtual recommendations and advice given by the DSS were compared with two independent experts (government doctors from the non-participating PHC centers).

Results: The overall percent agreement between the DSS and independent experts among 60 hypertensives on drug management was 85% (95% CI: 83.61-85.25). The kappa statistic for overall agreement for drug management was 0.659 (95% CI: 0.457-0.862) indicating a substantial degree of agreement beyond chance at an alpha fixed at 0.05 with 80% power. Receiver operator curve (ROC) showed a good accuracy for the DSS, wherein, the area under curve (AUC) was 0.848 (95% CI: 0.741-0.948). Sensitivity and specificity of the DSS were 83.33 and 85.71% respectively when compared with independent experts.

Conclusion: A point of care, pilot tested and validated DSS for management of hypertension has been developed in a resource constrained low and middle income setting and could contribute to improved management of hypertension at a primary health care level.

Show MeSH
Related in: MedlinePlus